Ratio Toolkit II: Liquidity, Solvency, Efficiency

Phase 2 — Financial Statement Analysis & Quality of Earnings · Week 19 · ~20 hours (one immersive week)

Two retailers. One holds three and a half rupees of current assets for every rupee of current bills; the other holds ninety-seven cents for every dollar. Both are among the most creditworthy retailers on earth.

No profitability ratio explains that. The profitability stack and DuPont, which came just before this, tell you whether a business earns well. They never tell you whether it survives and moves, and that's what the rest of the toolkit is for. Liquidity asks whether the company can pay this year's bills. Solvency and leverage ratios ask who really owns the balance sheet. Coverage asks whether earnings carry the interest and the rent, and efficiency asks how fast assets turn into cash. It all comes together in the Cash Conversion Cycle, the most story-rich number in analysis.

You'll also learn the mechanics of Enterprise Value and the market multiples (P/E, P/B, EV/EBITDA), so you can compute them now. What they mean for valuation waits for Phase 3. Behind you are the three statements with every note from Phase 1. Margins and DuPont are behind you too. Ahead, the returns-on-capital work builds ROIC on the working-capital and debt ideas you learn here, and the forensic weeks turn these same ratios into tripwires. These ratios serve Question 4 of the five-questions spine, what can kill it? Through efficiency they also serve Question 2, is it a good business?

Learning objectives

By the end you can:

1. Compute the current, quick, and cash ratios from any classified balance sheet, and interpret them against the business model rather than against a universal "2.0 is safe" rule.
2. Explain, and prove with numbers, why a current ratio below 1 and a negative working-capital cycle can be a sign of strength (Costco, HUL) or a sign of distress (a stretched industrial), and name the disclosures that separate the two.
3. Compute debt-to-equity under stated conventions, the DuPont financial-leverage multiplier, gross vs net debt, and net debt/EBITDA, and identify when netting cash against debt misleads (restricted, trapped, or operationally required cash).
4. Compute and rank the coverage family: interest coverage (TIE), EBITDA/interest, DSCR with principal, and fixed-charge coverage, including a lease-adjusted computation that works across US GAAP and IFRS/Ind AS reporters.
5. Compute all five turnover ratios (total asset, fixed asset, inventory, receivables, payables) with the correct numerators and denominators, convert each into days, and defend the conventions (COGS for inventory, purchases for payables, averages over ending balances).
6. Derive the Cash Conversion Cycle (CCC = DIO + DSO − DPO) from a cash timeline, compute it end-to-end for a real Indian and a real US company, and read it as a story: who finances whom in the value chain, and what growth will cost in cash.
7. Build Enterprise Value = market cap + debt + minority interest + preferred − cash from a filing, compute P/E, P/B, and EV/EBITDA mechanically, and state the consistency rule that pairs enterprise numerators with pre-interest denominators.
8. Sanity-check any of these ratios against sector norms ("what good looks like") and know where the entire toolkit does not apply (banks, NBFCs, insurers).

Prerequisites & connections

Five earlier modules feed this one. M1.03 gave you the classified balance sheet, where the current versus non-current split is the raw material of every liquidity ratio. M1.04 gave you the cash-flow statement, because coverage without cash flow is a half-truth. M1.06 covered receivables, inventory, and the cost-flow assumptions behind DIO and DSO. M1.08 covered debt at amortized cost, leases as debt, and the EBITDA comparability trap, all reused constantly here. M2.01 gave you margins and DuPont, and the financial-leverage multiplier taught there gets its full treatment below.

The ratios feed forward everywhere. M2.03 subtracts operating liabilities from operating assets to build invested capital, and the working-capital logic you learn here is half of that build. M2.04 (cash flow and FCF) explains why a rising CCC eats cash. M2.05 uses these ratios in time-series and peer benchmarking. M2.06–2.08 weaponize them: rising DSO and DIO against sales are the two most reliable shenanigan tripwires ever devised. M3.03 (cost of capital) builds synthetic credit ratings off the coverage ratios you learn here. M3.06 gives P/E, P/B, and EV/EBITDA their valuation meaning. Every Phase 5 sector playbook opens with the question this week trains: which of these ratios matters for this business model?

Core teaching content

4.0 Two questions, one toolkit

Profitability ratios ask whether the engine is any good. These ratios ask two different questions. First comes survival. A company dies when it can't pay a bill on the day the bill is due, whatever its margins are doing. Each of the survival ratios measures how far the company is from that day. Second comes velocity. Two companies with identical margins are different businesses if one turns its assets twice as fast. Efficiency ratios measure how quickly invested money comes back as cash.

Keep one distinction in front of you the whole way through: stock vs flow. Balance-sheet ratios such as the current ratio, D/E and net debt are stocks, a photograph of claims at one instant. Coverage ratios like TIE and DSCR are flows. They ask whether this year's earnings carry this year's obligations. An asset-rich, cash-poor company can look solvent as a stock while it's dying as a flow. A utility can carry heavy debt as a stock and service it without effort as a flow. Experts always read the pair together.

4.6 The Cash Conversion Cycle: the balance sheet as a clock

Now put the three day-counts together into a single number. Follow one rupee through Kaveri's operating cycle:

• Day 0. Kaveri buys raw material on credit. No cash has moved.
• Day 45 (DPO). Kaveri pays the supplier. Cash goes out.
• Day 60 (DIO). The finished goods sell, on credit. Still no cash in.
• Day 90 (DIO + DSO). The customer pays. Cash comes back.

From day 45 to day 90, a stretch of 45 days, Kaveri's own money is stuck in the machine. That gap is the Cash Conversion Cycle:

CCC = DIO + DSO − DPO = 60 + 30 − 45 = 45 days

The operating cycle, DIO + DSO = 90 days, is how long the machine takes end to end. Your supplier finances the first DPO days of it. The CCC is the stretch you have to finance with equity or debt. If the CCC is negative, you don't finance it with anybody's money, because cash comes in before it goes out. Everything else comes down to reading that number well.

Section 4.2 promised the numbers behind the DMart and Costco strategies. Here they are, end to end. They're illustrative, modeled on ~FY24/FY25 scale, so pull the actual filings and recompute. Ending balances keep it simple, so treat the outputs as approximate.

Read the table like an analyst, column by column. Both retailers spin inventory in about a month, which is elite grocery velocity and the ~11× turns from 4.2. Both collect from customers almost instantly, because they're cash-and-carry. So the whole strategic difference comes down to one line, DPO. Costco takes ~32 days of supplier credit. Suppliers finance the whole shelf, and the cycle nets to ~2 days, effectively zero. The CCC formula sees only inventory, receivables and payables, so it misses the ~$2.5bn of deferred membership fees. Add them and Costco's true cycle is negative. Then the machine funds itself and throws off a float. DMart pays suppliers in about a week, on purpose. It buys lower input prices with early payment and accepts a 29-day cycle financed by its own nearly debt-free balance sheet.

Neither number is "better" in the abstract. Each one is the balance-sheet shadow of a strategy, and seeing that is what reading working capital as a story means.

Every CCC is a sentence about power in a value chain, and it names who finances whom. Learn to say the sentence out loud:

• HUL (CCC deeply negative): suppliers finance HUL and distributors pre-pay, so the whole chain works for the brand owner. Its vendors are fragmented and its brands are must-stock.
• Costco (CCC ≈ 0, float-positive with memberships): suppliers and members finance the shelf between them, and Costco is the toll gate in the middle.
• DMart (CCC ≈ 29, by choice): shareholders finance the cycle. Suppliers are paid early in exchange for price, so DMart turns float it could have taken into a gross-margin advantage.
• A jeweller (CCC 80–150+): the company finances everyone, because gold waits for months before a customer buys. Titan's answer is to get customers to finance it through advance-purchase schemes, and suppliers through gold-on-lease. Worked example 2 computes it.
• A capital-goods/EPC firm (CCC often 60–120+ before advances): the company finances its customers through long receivables and milestone billing. So the sector lives and dies by customer advances, and the forensic receivable tripwires fire here first.

Operating WC scales with sales, so a growing company has to refill the cycle every year. Do the quick arithmetic. DMart's operating WC = 3,900 + 300 − 800 = ₹3,400 cr ≈ 6.8% of revenue. Grow revenue by ₹10,000 cr and roughly ₹680 cr of the resulting cash gets trapped in new inventory and receivables before a rupee reaches shareholders. Costco's WC is ~0.7% of revenue, and negative once memberships count, so growth costs it nothing. Historically, Amazon's negative cycle meant growth generated cash.

That one idea, ΔWC ≈ (WC% of sales) × Δrevenue, is the bridge to the free-cash-flow work coming up. It also explains how a rising CCC can strangle a fast grower while the P&L applauds.

The CCC works as a tripwire only on trend, never on level. A jeweller whose CCC has held at 120 for a decade is showing you its business model. An FMCG company drifting from 20 to 45 to 70 is a fire alarm. Inventory or receivables outrunning sales is the most reliable early signature of channel stuffing and demand rot. Later, the forensic work formalizes it, and Practice P10 drills it. Always break the move into its parts. A CCC that "improved" only because DPO lurched from 50 to 95 is the counterfeit float from 4.2, and calling it efficiency gets it backwards.

Common mistakes & how experts think differently

1. Grading every company against "current ratio 2.0, quick 1.0" is the novice move. The expert first asks how fast does this balance sheet spin? Costco at 0.97 is safer than a machine-tool maker at 2.5 whose inventory takes half a year to move and whose one big customer pays in 90 days. It's the business model that owns the threshold. It follows that a falling current ratio at a strengthening retailer can be good news, because it means more supplier financing. A rising one at an industrial can mean inventory piling up unsold.

2. Revenue in the inventory-turnover numerator imports the entire gross margin as phantom speed. Inventory is carried at cost while revenue includes the markup, so a 30% margin overstates turns by ~43%. Experts also check which convention a screener used before quoting its number, since screener.in and Bloomberg don't always agree with annual reports. DPO needs the same care. Use purchases if you can and COGS as a labeled fallback, with one convention across every company you compare.

3. Year-end balances for a fast grower, or for a seasonal business, mislead. Divide a full year's COGS by the March-31 inventory of a company that doubled and you understate the days money actually spent in the machine. A January year-end flatters a US retailer's inventory photo. Experts use averages and read quarterly balance sheets for seasonal names. When they're hunting shenanigans, they deliberately switch to ending balances and quarterly exit rates, because the freshest photograph shows the newest problems.

4. Netting every rupee of cash against debt ignores the cash that can't move. Restricted deposits like IndiGo's lessor collateral can't repay the parent's bonds tomorrow. Nor can escrowed project cash in RERA accounts or DSRA balances. Minimum operating cash can't either, and nor can cash trapped in foreign or partly owned subsidiaries. Experts read the cash note before computing net debt and can defend the split. Free cash nets, and the rest doesn't. Run the same skepticism the other way. A company whose "cash" is really inter-corporate deposits to promoter entities has no cash at all, and the forensic work comes back to it.

5. Leaving leases out, or counting them one and a half times, wrecks both ratios. Leave lease liabilities out of D/E and coverage and a Chipotle or an IndiGo looks absurdly safe. That error died with IFRS 16/Ind AS 116/ASC 842, which put the liability on every balance sheet. The error still alive today is the mismatch. One version puts lease-inclusive debt over a US GAAP rent-deducted EBITDA, which double-counts the lease. Another compares an Ind AS EBITDA straight across the border with a US GAAP one. Experts pick one lane and stay in it. That lane is EBITDAR with lease-inclusive obligations. Practice P8 drills it.

6. Reading negative working capital with one eye settles nothing, because the sign alone tells you nothing. HUL's float and a defaulting mid-cap's payables stretch carry identical signs. Before calling it strength, experts run the checklist from 4.2. They check DPO trend and relative power first. Then they read the MSME-interest and ageing disclosures and look for CFO confirmation. It works the other way too. A large positive CCC isn't automatically bad. For a jeweller it's the business model, and it's priced accordingly.

7. Coverage on EBIT alone, with rent nowhere, is what makes TIE flatter every rent-heavy business: retail, QSR, airlines, hotels, diagnostics chains. Experts reach for fixed-charge or EBITDAR coverage the moment the lease note is material. They also remember DSCR when principal amortizes. Bonds may bullet, but bank loans repay every year, and a company can cover interest 4× and still miss its schedule.

8. Multiples assembled from mismatched parts are everywhere. You'll see P/EBITDA, and EV over net income. You'll see EV missing the minority interest while EBITDA consolidates 100% of the subsidiary. There's trailing EPS at today's price set against a peer's forward EPS, and basic EPS when dilution is 5%. Experts run the consistency rule mechanically before any interpretation. That means matching the claim in the numerator to the income in the denominator. It also means using the same period and the same lease convention. Most "cheap" stocks that beginners find are arithmetic errors.

9. One ratio, one year, one company is the deepest difference in habit. Novices compute a number. Experts compute a panel: five to ten years, three to five peers, stock beside flow, level beside trend. Stock beside flow means net debt/EBITDA next to DSCR next to the maturity ladder. A single ratio is a word, and the panel is a sentence. The benchmarking work turns this instinct into a full method.

Worked examples

(All company figures illustrative, modeled on ~FY24/FY25 scale; pull the actual filings when you replicate. Synthetic companies are labeled.)

WE1: Chipotle completed: lease-adjusted coverage and leverage (US GAAP)

Section 4.4 left Chipotle halfway through. Here's the full picture, step by step:

Given (10-K scale): EBIT ~$1,920m, depreciation & amortization ~$340m, occupancy/rent cost ~$550m/yr, borrowings $0, interest on borrowings ~$0, operating-lease liabilities on the balance sheet (ASC 842) ~$4,600m, cash + investments ~$2,300m and shareholders' equity ~$3,700m.

1. Naive reads. TIE = 1,920 ÷ 0 is undefined (infinite). D/E = 0 ÷ 3,700 = 0. Debt-free fortress, says the naive toolkit.
2. EBITDA and EBITDAR. EBITDA = EBIT + D&A = 1,920 + 340 = $2,260m. Under US GAAP, rent was deducted inside opex, so add it back for the lease-neutral earnings base: EBITDAR = 2,260 + 550 = $2,810m.
3. Fixed-charge coverage (EBIT convention). FCC = (EBIT + rent) ÷ (interest + rent) = (1,920 + 550) ÷ (0 + 550) = 2,470 ÷ 550 = 4.5×.
4. EBITDAR coverage (lender convention). EBITDAR ÷ (interest + lease payments) = 2,810 ÷ 550 = 5.1×. Both say the same thing. It's healthy, and it's decidedly finite.
5. Lease-adjusted leverage. Adjusted debt = lease liabilities = $4,600m. Cross-check it with the pre-IFRS-16 rating-agency heuristic, 8 × annual rent = 8 × 550 = $4,400m. The on-balance-sheet number and the old shortcut agree within ~5%. Adjusted net debt = 4,600 − 2,300 = $2,300m. Adjusted net debt ÷ EBITDAR = 2,300 ÷ 2,810 = 0.8×. Lease-adjusted D/E = 4,600 ÷ 3,700 = 1.2×.
6. Verdict. "Zero debt" was presentation. In economic terms Chipotle is a moderately committed credit at 1.2× lease-adjusted D/E, ~0.8× net leverage and ~5× fixed-charge cover. That's excellent, and you know it because you measured it. Any rent-heavy business you ever analyze gets the same treatment.

WE2: Titan: a long CCC and the art of making others finance it (India)

The jeweller is the anti-Costco, because gold stays in display cases for months. Given (consolidated, illustrative ~FY24 scale, ₹ cr): revenue 51,000, COGS 38,800, inventory 16,000, trade receivables 1,300, trade payables + gold-on-lease obligations 8,000, and customer advances of ~3,000, including jewellery purchase schemes.

1. DIO = 16,000 ÷ 38,800 × 365 = 150.5 days. Five months of metal and stones on display.
2. DSO = 1,300 ÷ 51,000 × 365 = 9.3 days. Retail customers pay at the counter, so the small receivable is mostly institutional and watches.
3. DPO = 8,000 ÷ 38,800 × 365 = 75.3 days, and the trick is buried in it. A large slice is gold on lease, a jewellery-trade arrangement where banks lend the metal itself and Titan pays for it only around the time it sells. That's supplier financing built for a 150-day shelf.
4. CCC = 150.5 + 9.3 − 75.3 = 84.5 days, call it ~85 days that someone must finance.
5. Who finances whom, quantified. Operating WC = 16,000 + 1,300 − 8,000 = ₹9,300 cr ≈ 18.2% of revenue. Growing revenue by ₹5,000 cr therefore traps roughly 0.182 × 5,000 ≈ ₹910 cr of new cash, before you count the offsets. And the offsets are the strategy. There's ~₹3,000 cr of customer advances, from schemes where buyers deposit monthly toward future jewellery. That way customers finance the display case, which is the Costco-membership move in Indian dress. Then there's the gold-on-lease line. The CCC formula sees neither the advances nor their float, but the operating-WC view does. Always read both.
6. Verdict. An 85-day cycle is the price of the category, and Titan manages it down by borrowing the shelf through gold lease and pre-selling it through schemes. Now picture a weaker jeweller with the same DIO but no advances. Its payables are stretched by leaving artisans unpaid. Same arithmetic, opposite creditworthiness. The story test from 4.2 decides it. A number alone can't.

WE3: Meridian Tools (synthetic): the counterfeit float, caught by the checklist

A mid-cap industrial shows "improving" working capital. Here are three years of data in ₹ cr. COGS is flat at 1,825, which works out to ₹5 cr a day and makes the days easy to see.

1. Days. DIO = 440/5 = 88, which rises to 104 before reaching 126. DSO = 375 ÷ 2,500 × 365 = 54.8, then 63.7, then 78.4. DPO = 275/5 = 55, then 70, then 95.
2. CCC. 88 + 54.8 − 55 = 87.8, then 97.7, then 109.4. Despite a 40-day payables stretch, the cycle lengthened by 22 days. Inventory and receivables are ballooning faster than suppliers can be made to fund them.
3. Checklist from 4.2. (i) DPO lurching from 55 to 95 while revenue is flat is a distress signature. Power has nothing to do with it. (ii) On the power test, a mid-cap tool maker doesn't dictate 95-day terms to steel suppliers. (iii) India's tell is there too. MSME interest accrued going from 0 to 2 to 11 means small vendors are being paid beyond the MSMED Act's 45 days. That's stretch in plain sight, in the notes. (iv) Cash confirms it. CFO/PAT = 210/160 = 1.31, then 0.52, then −0.43. The profits are on paper, and the cycle is eating cash.
4. Verdict. Every flag fires. This is a company borrowing from its suppliers because banks have stopped lending, the counterfeit version of float. With PAT barely moving, the P&L would never tell you. The working-capital panel shouts it two years before the default. You'll hunt this same pattern in the forensic weeks.

WE4: Bharti Airtel: a full EV build with minority interest (India)

Here's why NCI matters in real life. Figures are illustrative, at ~FY24/FY25 scale, in ₹ cr, and every input moves, so pull current data. Market cap is ~9,00,000, which is price × shares on the day you compute. Borrowings are ~1,55,000. Lease liabilities are ~65,000. These cover towers plus spectrum-adjacent leases, since a telecom is lease-heavy by nature. Cash + investments are ~15,000. Non-controlling interest at book is ~35,000, mostly outside shareholders of the listed Airtel Africa subsidiary. Preferred is nil. Consolidated EBITDA is ~80,000.

1. EV = 9,00,000 + (1,55,000 + 65,000) + 35,000 + 0 − 15,000 = ₹11,40,000 cr.
2. EV/EBITDA = 11,40,000 ÷ 80,000 = 14.3×.
3. What each adjustment did. Market cap ÷ EBITDA alone = 11.3×. That understates the price of the business by three full turns. Lenders and lessors own a quarter of this enterprise, and the consolidated EBITDA includes Airtel Africa profits that partly belong to its minority shareholders. Skip the NCI and you're buying 100% of the EBITDA with ~96% of the claims. That's a free lunch, which means it's an error.
4. Refinement to file away. Book NCI understates when the subsidiary trades publicly. The better move is the market value of the minority stake, which is Airtel Africa's London-listed price × outside ownership. For now you only need the mechanics. The valuation work handles the judgment calls, including whether to use forward EBITDA and how to read 14× against peers.

WE5: Sahyadri Pharma (synthetic): a long, receivables-heavy cycle built on average balances (India)

The generics maker is in the opposite corner of the sector table from grocery. Its cycle is long, and receivables do most of the damage. Titan in WE2 used ending balances. This one uses two balance sheets and average balances, the convention 4.5 insists on. Given (consolidated, ₹ cr): revenue 8,000 and COGS 4,400, a 45% gross margin, which is decent for generics. Inventory opens at 1,150 and closes at 1,250. Trade receivables open at 1,700 and close at 1,900. Trade payables open at 700 and close at 780.

1. Averages and purchases first. Average inventory = (1,150 + 1,250) ÷ 2 = 1,200, average receivables = 1,800 and average payables = 740. Purchases = COGS + closing − opening inventory = 4,400 + (1,250 − 1,150) = 4,500.
2. DIO. 365 × 1,200 ÷ 4,400 = 99.5 days. Three-plus months of active ingredient and finished stock, with batch testing and regulatory shelf-life baked in.
3. DSO. 365 × 1,800 ÷ 8,000 = 82.1 days. Hospitals and government tenders pay slowly, and so do export distributors. That's where a generics maker's cash goes to sleep.
4. DPO. On purchases, 365 × 740 ÷ 4,500 = 60.0 days. The COGS proxy gives 365 × 740 ÷ 4,400 = 61.4, which is close because inventory barely moved.
5. CCC. 99.5 + 82.1 − 60.0 = 121.6 days. The operating cycle, DIO + DSO, is 181.7 days end to end. Suppliers finance 60 of it, and the company funds ~122 days itself, squarely inside the sector table's 90-to-150-day pharma band. Receivables, at 82 days, are the largest block, bigger than inventory. So the lever here is collections. Your warehouse matters less.
6. What growth costs in cash. Operating WC = 1,250 + 1,900 − 780 = ₹2,370 cr ≈ 29.6% of revenue. Grow revenue by ₹2,000 cr and roughly 0.296 × 2,000 ≈ ₹593 cr is swallowed by new stock and unpaid invoices before a rupee reaches shareholders. Nearly a third of every extra sale gets trapped. That's how a fast-growing generics firm manages to report rising profit and falling cash at the same time, as the cash-flow work will show.
7. Verdict. Nothing about a 122-day cycle signals distress in this category, and DSO dominates it. Its fixes are tighter collections and receivables factoring, never inventory cuts. Set this beside the ~2-to-29-day DMart and Costco cycles of 4.6. The three formulas are the same, but the balance-sheet clock is completely different, because the business model is. Read the CCC, then name which of the three clocks is loudest.

WE6: Norse Steel Corp (synthetic): the cyclical leverage ladder, and what a downturn does to it (US)

Section 4.3 asserted that a steel company at the top of the cycle is a time bomb "because the denominator halves in a downturn while the numerator doesn't." Here's that sentence turned into arithmetic. You'll build the full coverage-and-leverage ladder at mid-cycle, then rebuild it on the identical balance sheet after EBITDA halves. Given (top of the cycle, $m): EBIT 1,100, D&A 500 (so EBITDA 1,600), interest 200, total debt including leases 4,600, cash 600 (all free), scheduled principal due this year 300, cash taxes 150 and maintenance capex 250.

1. The peak-cycle ladder, bluntest to sharpest. TIE = EBIT ÷ interest = 1,100 ÷ 200 = 5.5×. EBITDA/interest = 1,600 ÷ 200 = 8.0×. Net debt = 4,600 − 600 = 4,000, so net debt/EBITDA = 4,000 ÷ 1,600 = 2.5×, which is the "managed leverage" band in the 4.3 table. CFADS = 1,600 − 150 − 250 = 1,200 and service = 200 + 300 = 500, so DSCR = 1,200 ÷ 500 = 2.4×. Every rung says "comfortable", and a naive analyst files this as a safe credit.
2. Now halve EBITDA (an ordinary steel downturn). Price and volume fall together. EBITDA drops to 800 and EBIT to 800 − 500 = 300. Debt hasn't moved, because falling margins don't repay 4,600 of bonds and leases.
3. The ladder after the fall. TIE = 300 ÷ 200 = 1.5×, down from 5.5× to a distress conversation. EBITDA/interest = 800 ÷ 200 = 4.0×. And net debt/EBITDA = 4,000 ÷ 800 = 5.0×. The ratio doubled, straight through covenant territory and into high-yield/LBO territory above 4 to 4.5×, only because the denominator halved while the numerator stood still.
4. DSCR, the sharpest read. In the downturn cash taxes shrink to ~25, but maintenance capex can't be deferred forever, so it stays at ~250. CFADS = 800 − 25 − 250 = 525 and DSCR = 525 ÷ 500 = 1.05×. The plant can barely pay its own schedule from its own cash. One more turn down and it refinances or defaults.
5. Verdict. The identical balance sheet reads "2.5× and comfortable" at the top and "5.0× and covenant-breaching" at the bottom. Nothing about the company changed except the weather. So 4.3 calibrates leverage to EBITDA stability, and a cyclical has to be judged on mid-cycle earnings, never on peak earnings. A regulated utility at 3.5× can be a safer credit than this steelmaker at 2.5×.

Practice set

All the companies are synthetic, with figures in ₹ cr unless stated. P1 to P3 are guided, so follow the steps. P4 to P8 are independent, and P9 to P12 are timed. Set a clock for those, because they train exam speed. Each problem is followed straight away by its full solution. Work it first, then check.

P1 (guided). The liquidity trio plus one. Arjun Fasteners' current assets: cash 120 (of which margin money pledged with banks: 30), marketable securities 80, trade receivables 300, inventory 500, prepaid expenses 40. Current liabilities: trade payables 350, current maturities of long-term debt 150, accrued expenses 100, short-term borrowings 50. Annual cash operating expenses (excl. D&A) 3,650. Compute the current, quick, and cash ratios and the defensive interval. Then recompute the quick ratio counting only unrestricted cash. Which single current-liability line deserves the most attention, and why?

Solution. CA = 120 + 80 + 300 + 500 + 40 = 1,040. CL = 350 + 150 + 100 + 50 = 650. Current = 1,040 ÷ 650 = 1.60. Quick = (120 + 80 + 300) ÷ 650 = 500 ÷ 650 = 0.77. Cash ratio = (120 + 80) ÷ 650 = 200 ÷ 650 = 0.31. Daily cash opex = 3,650 ÷ 365 = 10, so defensive interval = 500 ÷ 10 = 50 days of operation if all inflows stopped. Strict quick = (90 + 80 + 300) ÷ 650 = 470 ÷ 650 = 0.72, because the pledged 30 is collateral, and you can't fire collateral as ammunition. The line to watch is the current maturities of 150, a hard calendar obligation. Cash + securities (200, or 170 free) can't cover it plus the 50 of short-term borrowings without relying on collections. So the refinancing and rollover plan in the borrowings note is required reading.

P2 (guided). Leverage under stated conventions. Deccan Cables: borrowings 1,800 (of which current maturities 300), lease liabilities 400, trade payables 900, other non-financial liabilities 300, shareholders' equity 2,200. Cash 500 (of which 150 restricted), liquid mutual funds 200. EBITDA 700. Compute (a) financial D/E, (b) total liabilities/equity, (c) net debt the naive way and the right way, (d) net debt/EBITDA and its band from 4.3.

Solution. (a) Financial debt = 1,800 + 400 = 2,200. Current maturities are already inside the 1,800, so don't count them twice. D/E = 2,200 ÷ 2,200 = 1.0×. (b) Total liabilities = 1,800 + 400 + 900 + 300 = 3,400, giving 3,400 ÷ 2,200 = 1.55×. It's bigger, but it counts payables that carry no interest, so say which convention you're quoting. (c) Naive net debt = 2,200 − 500 − 200 = 1,500. Done right, only free cash nets, giving 2,200 − (500 − 150) − 200 = 1,650. (d) 1,650 ÷ 700 = 2.4×, the "managed leverage" band. That's fine for stable cash flows and worth stress-testing for a cyclical. The naive version, 2.1×, flattered by a quarter turn.

P3 (guided). Days done right. A trading company: revenue 9,125, COGS 6,570. Inventory: opening 900, closing 1,080. Receivables: opening 700, closing 800. Payables: opening 540, closing 600. Compute DIO (correct and with the revenue-numerator error), DSO, DPO on purchases and on the COGS proxy, and the CCC.

Solution. The averages are inventory 990, receivables 750, payables 570. DIO = 365 × 990 ÷ 6,570 = 55.0 days. Done the wrong way, 365 × 990 ÷ 9,125 = 39.6 days, the 28% gross margin masquerading as 15 days of speed. DSO = 365 × 750 ÷ 9,125 = 30.0 days. Purchases = COGS + closing inv − opening inv = 6,570 + 180 = 6,750, so DPO = 365 × 570 ÷ 6,750 = 30.8 days. The COGS proxy gives 365 × 570 ÷ 6,570 = 31.7, close here because the inventory build was small. CCC = 55.0 + 30.0 − 30.8 = 54.2 days.

P4. Float or stretch? Two companies each show operating working capital of −400. Co A's DPO went 88, 90, 91 over three years while revenue rose 12%/yr. Its CFO was 850 against PAT of 700. There was no MSME interest, and 96% of payables were aged under 90 days. Co B's DPO went 52, 68, 95 on flat revenue. Its CFO was 60 against PAT of 350. MSME interest accrued was 18 and rising, and the ageing schedule shows a growing >1-year bucket. Classify each and cite the evidence in checklist order.

Solution. A is real float. Its DPO is stable, which points to payment terms and away from payment failure. The float grows with double-digit revenue growth. CFO/PAT of 1.21 confirms the cash is real, and the ageing and MSME disclosures are clean too. B is counterfeit, a distress stretch. DPO lurching +43 days on flat revenue means suppliers are financing it against their will. CFO/PAT of 0.17 says the profits aren't cash. Statutory MSME interest and an aging >1-year bucket are evidence of unpaid vendors, right there in the notes. Same −400, opposite conclusions. The sign of working capital is where analysis starts, never where it ends.

P5. The coverage family, complete. Nilgiri Power: EBIT 640, D&A 260, interest 200, scheduled principal repayment this year 250, cash taxes 90, maintenance capex 120. Lender covenant: DSCR ≥ 1.3× on a CFADS basis. Compute TIE, EBITDA/interest, EBITDA-DSCR, CFADS-DSCR, and the % fall in EBITDA the covenant can absorb.

Solution. TIE = 640 ÷ 200 = 3.2×, so leverage is a live constraint. EBITDA = 640 + 260 = 900, which gives EBITDA/interest = 4.5×. Debt service = 200 + 250 = 450, so EBITDA-DSCR = 900 ÷ 450 = 2.0×. CFADS = 900 − 90 − 120 = 690, and CFADS-DSCR = 690 ÷ 450 = 1.53×, which passes. Now the headroom. Your covenant needs CFADS = 1.3 × 450 = 585, which means EBITDA of 585 + 90 + 120 = 795. The cushion is (900 − 795) ÷ 900 = 11.7%. A power producer with take-or-pay contracts can live with that. A merchant plant can't. Notice what each refinement did. Adding principal to interest cut the ratio. Taking out taxes and capex cut it again, until it was nearly half where it started.

P6. Rent is a fixed charge. Prairie Home Stores (US GAAP): EBIT 480, interest 60, rent 240. Compute TIE and fixed-charge coverage; reconcile the two verdicts.

Solution. TIE = 480 ÷ 60 = 8.0×, which looks unconstrained. FCC = (480 + 240) ÷ (60 + 240) = 720 ÷ 300 = 2.4×, which is moderately committed. To reconcile them, remember that rent was deducted before EBIT and interest wasn't. So TIE compares a post-rent profit with a rent-free obligation line. Put rent back on both sides and the landlord is the biggest "lender." For any retailer, restaurant, airline or hotel, trust FCC or EBITDAR coverage. TIE is decoration.

P7. Averages vs ending: the growth distortion. A company grew revenue from 600 to 1,200 during the year. Receivables: opening 100, closing 300. Compute DSO on average and on ending balances. Explain why they diverge and when you would deliberately prefer each.

Solution. Average receivables of 200 give DSO = 365 × 200 ÷ 1,200 = 60.8 days. Ending receivables give 365 × 300 ÷ 1,200 = 91.3 days. They diverge because most of the year's revenue happened at the old scale, while the closing balance reflects the new one. Averages match the flow to the stock. So prefer averages for fair efficiency measurement and peer comparison. When you're screening for deterioration, prefer ending balances. Better still, use Q4 annualized, 365 × 300 ÷ (Q4 revenue × 4). The freshest photo shows the newest behavior, and a stuffing episode in March vanishes inside a full-year average.

P8. The EBITDAR twins (promised in 4.4). Twin U reports under US GAAP with its store leases as operating leases. Twin I reports under Ind AS 116. Their economics are identical: revenue 2,000, cash opex other than rent 1,580, annual lease payment 120 (interest portion 25, principal portion 95), owned-asset depreciation 80, borrowings 600 at interest 30, cash 100. Both carry a lease liability of 500, since ASC 842 puts operating-lease liabilities on the balance sheet too. What differs is P&L geography. Twin I's ROU depreciation is 95. Compute each twin's EBITDA, EBIT, naive TIE, and naive net debt/EBITDA, then repair both with the EBITDAR convention and show they converge.

Solution. Twin U has rent in opex. EBITDA = 2,000 − 1,580 − 120 = 300, and EBIT = 300 − 80 = 220. TIE = 220 ÷ 30 = 7.3×. Net debt/EBITDA = (600 − 100) ÷ 300 = 1.67×, treating only borrowings as debt, the way a naive analyst would. Twin I has no rent line. EBITDA = 2,000 − 1,580 = 420, and EBIT = 420 − 80 − 95 = 245. Interest = 30 + 25 = 55, so TIE = 245 ÷ 55 = 4.5×. Net debt/EBITDA = (600 + 500 − 100) ÷ 420 = 2.38×. Identical twins get four different numbers, with nothing behind them but accounting geography.

Now the repair. EBITDAR_U = 300 + 120 = 420 = EBITDA_I, because Ind AS EBITDA already excludes lease costs. Lease-inclusive net debt for both = 600 + 500 − 100 = 1,000, so 1,000 ÷ 420 = 2.38× for both. Coverage = EBITDAR ÷ (interest on borrowings + total lease payment) = 420 ÷ (30 + 120) = 2.8× for both. Converged. The lesson fits in one line: pick the lease-inclusive lane and every border disappears.

P9 (timed, 10 min). EV and the multiple set. Sundaram Foods: share price ₹450, shares 20 cr, borrowings 1,200, lease liabilities 300, NCI (book) 250, preferred nil, cash 700, liquid investments 300. EBITDA 750, PAT attributable to parent 500, common equity 2,500. Compute EV, EV/EBITDA, P/E, and P/B, and name the error a careless analyst makes with PAT here.

Solution. Market cap = 450 × 20 = 9,000. EV = 9,000 + 1,200 + 300 + 250 − (700 + 300) = 9,750. EV/EBITDA = 9,750 ÷ 750 = 13.0×. EPS = 500 ÷ 20 = 25, so P/E = 450 ÷ 25 = 18.0×. BVPS = 2,500 ÷ 20 = 125, so P/B = 450 ÷ 125 = 3.6×. The careless error is using consolidated PAT including the minority's share in EPS. The numerator of P/E must be profit attributable to the parent's shareholders only, just as EV must add the NCI claim when EBITDA consolidates it.

P10 (timed, 10 min). Read the panel. Three years: revenue 100, 110, 121. DSO 45, 58, 74. DIO 60, 71, 85. DPO 50, 51, 52. CFO/NI 0.95, 0.60, 0.20. Compute the CCC path and write the three-sentence diagnosis an expert would.

Solution. CCC = 45+60−50 = 55, then 58+71−51 = 78, then 74+85−52 = 107. Here's the diagnosis. Reported sales grow 10% a year. Receivables and inventory grow far faster. Goods are being pushed into the channel and invoices booked that customers aren't paying on time, the classic pre-revenue-shenanigan signature the forensic work studies. A cash conversion cycle that nearly doubles while DPO stays flat means every rupee of reported growth is being bought with trapped cash, and CFO/NI collapsing from 0.95 to 0.20 confirms it. Until proven otherwise, treat reported earnings as unfinanced claims that haven't turned into cash, and expect a receivables write-off or a "channel correction" within eighteen months.

P11 (timed, 5 min). Who finances whom: match and justify. Company W: DIO 30, DSO 3, DPO 30. Company X: DIO 2, DSO 5, DPO 60. Company Y: DIO 160, DSO 10, DPO 70. Company Z: DIO 45, DSO 90, DPO 55. The four business models: B2B capital equipment; grocery retail; jeweller; e-commerce marketplace selling from consignment-light inventory. Match them and compute each CCC. Then say who finances whom in one phrase each.

Solution. W is grocery retail. CCC = 30+3−30 = 3, so suppliers finance almost the whole shelf and customers pay cash. X is the e-commerce marketplace. CCC = 2+5−60 = −53. It collects from buyers instantly and holds days of inventory. It pays sellers or vendors on 60-day cycles, so everyone else finances the platform. Y is the jeweller. CCC = 160+10−70 = 100. Here the company and its gold lessors finance a slow shelf. Z is capital equipment. CCC = 45+90−55 = 80, dominated by DSO. The company finances its customers. That's why order advances and milestone billing decide survival here.

P12 (timed, 5 min). Where the toolkit stops. A screener shows HDFC Bank with "current ratio 0.4, D/E 8.2×, interest coverage 1.4×" and flags it deep red on all three. Write the four-line correction you'd give the screener's author.

Solution. (1) Banks don't publish classified balance sheets, so a bank "current ratio" is computed on categories that don't exist. A bank's liquidity is measured by LCR, ALM buckets and CASA. (2) D/E of 8 to 11× is the business model, because banks are levered spread businesses. The solvency question is regulatory capital, CRAR/CET1 against Basel/RBI minimums, and corporate D/E bands don't answer it. (3) Interest expense is a bank's raw-material cost, so interest coverage is a category error. NIM is the margin metric. (4) The general rule is to classify the business model first, then pick the ratio set. These ratios are built for non-financial companies, and Phase 5 carries the banking instruments.

P13 (guided). The full turnover panel. Konkan Ceramics (synthetic; ₹ cr). Two balance sheets give: revenue 6,000; COGS 4,200; average total assets 4,000; average net PP&E 1,500; average inventory 700 (opening 650, closing 750); average receivables 500; average payables 430. Compute all five turnover ratios, convert inventory, receivables, and payables to days (purchases-based DPO), and the CCC.

Solution. Purchases = COGS + closing − opening inventory = 4,200 + (750 − 650) = 4,300. Total asset turnover = 6,000 ÷ 4,000 = 1.5×, so each rupee of assets throws off ₹1.50 of sales a year. Fixed asset turnover = 6,000 ÷ 1,500 = 4.0×. Inventory turnover = COGS ÷ avg inventory = 4,200 ÷ 700 = 6.0×, which gives DIO = 365 ÷ 6.0 = 60.8 days. Receivables turnover = 6,000 ÷ 500 = 12.0×, which gives DSO = 30.4 days. Payables turnover = purchases ÷ avg payables = 4,300 ÷ 430 = 10.0×, which gives DPO = 36.5 days. CCC = 60.8 + 30.4 − 36.5 = 54.7 days. The two turnovers that never become days are still doing work. Total-asset and fixed-asset turnover are the turnover half of DuPont, the lever that converts a thin margin into a real ROIC.

P14 (guided). The cyclical leverage trap, in numbers. Vindhya Cement (synthetic; ₹ cr), at the top of its cycle: EBIT 900, D&A 400, interest 150, total debt incl. leases 3,200, free cash 400. (a) Compute TIE and net debt/EBITDA and name the 4.3 band. (b) A downturn halves EBITDA; recompute both. (c) In one line, why did net debt/EBITDA move so far?

Solution. (a) EBITDA = 900 + 400 = 1,300 and net debt = 3,200 − 400 = 2,800. TIE = 900 ÷ 150 = 6.0× and net debt/EBITDA = 2,800 ÷ 1,300 = 2.15×, which is "managed leverage (2–3×)." (b) EBITDA halves to 650, so EBIT = 650 − 400 = 250. Now TIE = 250 ÷ 150 = 1.67×, a distress conversation, and net debt/EBITDA = 2,800 ÷ 650 = 4.31×, which is "high-yield/LBO territory (>4×)." (c) Debt is a fixed stock that a bad year doesn't shrink. When the denominator halves, the ratio roughly doubles. That's why a cyclical is judged on mid-cycle earnings, never the peak.

P15. EV when the company is net cash. Ashwa Technologies (synthetic IT services, ₹ cr): share price ₹1,200 and 40 cr shares, borrowings 0, lease liabilities 2,000, NCI 0, preferred 0, cash + liquid investments 18,000. EBITDA 3,000, PAT (parent) 2,400, common equity 12,000. Compute EV, EV/EBITDA, P/E, and P/B, state why EV sits below market cap, and name one multiple you must not form.

Solution. Market cap = 1,200 × 40 = 48,000. EV = 48,000 + 0 + 2,000 + 0 + 0 − 18,000 = 32,000. EV/EBITDA = 32,000 ÷ 3,000 = 10.7×. EPS = 2,400 ÷ 40 = ₹60, so P/E = 1,200 ÷ 60 = 20.0×. BVPS = 12,000 ÷ 40 = ₹300, so P/B = 1,200 ÷ 300 = 4.0×. EV is ₹16,000 cr below market cap because cash and investments of 18,000 dwarf the only debt, 2,000 of leases. That's the Infosys mirror from 4.7. A net-cash company's business is worth less than its equity, because the buyer pockets the till. The multiple you must never form is P/EBITDA. Price is an equity-only claim while EBITDA belongs to all capital providers, so use EV/EBITDA.

P16. Rent and preferred are both fixed charges. Beacon Retail Inc. (synthetic, US GAAP, $m): EBIT 620, rent 300, interest 80, preferred dividends 40. Compute TIE, fixed-charge coverage on rent alone, and fixed-charge coverage adding preferred dividends. Then say which verdict a lender trusts.

Solution. TIE = 620 ÷ 80 = 7.75×. It looks unconstrained because rent of 300 was deducted above EBIT while the small interest line wasn't. FCC (rent) = (620 + 300) ÷ (80 + 300) = 920 ÷ 380 = 2.42×. Adding preferred dividends as a fixed charge gives (620 + 300) ÷ (80 + 300 + 40) = 920 ÷ 420 = 2.19×. A fuller treatment grosses preferred dividends up to pre-tax, since they're paid after tax, so this simplification understates the burden slightly. A lender trusts the ~2.2 to 2.4× fixed-charge number over the 7.75× TIE. For a rent-heavy retailer the landlord and the preferred holders are lenders in all but name.

P17. What growth costs each business in cash. Two synthetic firms (₹ cr). Co G has inventory 1,200, receivables 800, payables 500 and revenue 10,000. Co H, an FMCG company, has inventory 900, receivables 300, payables 2,400 and revenue 12,000. For each, compute operating working capital and WC as a % of sales. Then work out the cash trapped or released if revenue grows ₹3,000 cr, and say which business self-funds its growth.

Solution. For Co G, operating WC = 1,200 + 800 − 500 = 1,500, which is 15.0% of revenue. So ΔWC ≈ 0.15 × 3,000 = ₹450 cr trapped before a rupee reaches shareholders. For Co H, operating WC = 900 + 300 − 2,400 = −1,200, which is −10.0% of revenue, so ΔWC ≈ −0.10 × 3,000 = −₹300 cr — growth releases ₹300 cr of cash. Co H self-funds its growth on suppliers' and customers' money, and in fact over-funds it. That's the HUL/Amazon pattern. Co G has to find ₹450 cr of retained or external cash for every ₹3,000 cr of new sales. The sign of operating WC decides whether growth is a cash source or a cash sink, and that's the bridge to the free-cash-flow work coming next.

P18 (timed, 12 min, stretch). The whole toolkit on one company. Cobalt Industrial Inc. (synthetic, US GAAP, $m). Current assets: cash 300, receivables 600, inventory 900, prepaids 100. Current liabilities: trade payables 700, current maturities of long-term debt 200, accrued 150, short-term borrowings 100. Borrowings 1,400 (incl. the 200 of current maturities), operating-lease liabilities 600. Revenue 5,000, COGS 3,650, EBITDA (as reported, US GAAP) 700, annual rent 180, interest 90. Share price $50, 100m shares. Use ending balances, and note that you did. Compute (a) current and quick ratios, (b) net debt/EBITDA on the lease-consistent base, (c) EBITDAR coverage, (d) CCC with a COGS-proxy DPO and (e) lease-consistent EV/EBITDA. Then write a two-line verdict.

Solution. (a) CA = 300 + 600 + 900 + 100 = 1,900 and CL = 700 + 200 + 150 + 100 = 1,150, so current = 1,900 ÷ 1,150 = 1.65× and quick = (300 + 600) ÷ 1,150 = 0.78×. (b) Total debt incl. leases = 1,400 + 600 = 2,000, and net debt = 2,000 − 300 = 1,700. Lease-inclusive debt demands a lease-inclusive EBITDA, so use EBITDAR = 700 + 180 = 880. That gives net debt/EBITDAR = 1,700 ÷ 880 = 1.93×. The trap is 1,700 ÷ 700 rent-deducted EBITDA = 2.43×, which double-counts the lease. (c) EBITDAR coverage = 880 ÷ (interest + lease payments) = 880 ÷ (90 + 180) = 3.26×. (d) DIO = 365 × 900 ÷ 3,650 = 90.0, DSO = 365 × 600 ÷ 5,000 = 43.8 and DPO = 365 × 700 ÷ 3,650 = 70.0, labeled as the COGS proxy. CCC = 90.0 + 43.8 − 70.0 = 63.8 days. (e) Market cap = 50 × 100 = 5,000, so EV = 5,000 + 2,000 − 300 = 6,700 and EV/EBITDAR = 6,700 ÷ 880 = 7.6×, with lease-inclusive EV paired with lease-inclusive EBITDA.

Verdict: a moderately leveraged industrial, at ~1.9× lease-adjusted net debt and ~3.3× fixed-charge cover, with a ~64-day cycle typical of a maker that finances some customer credit. Nothing alarms. Still, current maturities of 200 plus short-term borrowings of 100, against only 300 of cash, is the refinancing line to watch. As for scoring, five parts in 12 minutes counts as fluent. The lease-consistency step in (b) and (e) is the one most people get wrong.

Applied mini-project

You've seen DMart and Costco through illustrative numbers. Now tell the working-capital story of both retailers from the actual filings, and find out where reality complicates the clean version.

Pull two sets of filings. (1) Avenue Supermarts' latest annual report, from the investor-relations page at avenuesupermarts.com or the BSE/NSE filings portal. You need the consolidated balance sheet, P&L and cash-flow statement, plus the notes on borrowings, trade payables (incl. the MSME/ageing disclosures) and leases. (2) Costco's latest 10-K from SEC EDGAR (sec.gov → company search "Costco" → 10-K). You need the same sections plus the membership-fee (deferred revenue) figure and the leases note. Take three fiscal years of each, using the filings' comparatives plus one prior filing.

Build (a spreadsheet, one tab per company):

1. Liquidity: current, quick, cash ratios, each year.
2. Leverage: financial D/E including leases, and net debt, stating how you treat any restricted cash. Then net debt/EBITDA on a lease-consistent basis. Costco reports under US GAAP and DMart under Ind AS 116, so apply the P8 repair.
3. Coverage: TIE and EBITDAR ÷ (interest + lease payments).
4. All five turnovers with days, on average balances. Then the three day-counts combined into the CCC, each year.
5. EV as of today's prices (state the date), then EV/EBITDA, P/E, P/B.
6. Memo (one page): the who-finances-whom story of each company, the strategy you can read off the DPO line, and what the CCC trend over three years says. Add one place where the real filing surprised you compared with the illustrative version above. There will be one, so find it.

Rubric (score yourself without mercy; 0 = missing/wrong, 1 = partial, 2 = correct and defended):

Pass: ≥ 9/12, with no zero on any criterion. Keep the spreadsheet. It becomes the seed of your benchmarking template next, and the Phase 2 milestone of hand-spread filings expects this workbook format.

Case Lab: run case CL-2.2 in the app's Cases tab for this module. It is a synthetic manufacturer whose headline CCC "improves" year on year while the notes quietly fill with MSME delayed-payment interest, walked through the float-versus-stretch checklist, the coverage trio, and the net-debt/EBITDA band end to end.

Reading & resources

Core texts (in the order to reach for them):

• Benjamin Graham & Spencer Meredith, The Interpretation of Financial Statements. The 100-page classic; the ratio-definition entries in Part II. [Paid] [Beginner]
• Pat Dorsey, The Five Rules for Successful Stock Investing. Ch. 5 ("Financial Statements Explained") and Ch. 6 ("Analyzing a Company — The Basics"): the practitioner's version of this toolkit. [Paid] [Beginner]
• K. R. Subramanyam, Financial Statement Analysis (11th ed.). The credit-analysis chapter (Ch. 10 in recent editions): liquidity, working capital, solvency, and coverage with full rigor. [Paid] [Intermediate]
• Fridson & Alvarez, Financial Statement Analysis: A Practitioner's Guide (5th ed.). The credit-analysis and ratio chapters; how debt investors actually read these numbers. [Paid] [Advanced]
• White, Sondhi & Fried, The Analysis and Use of Financial Statements (3rd ed.), Ch. 4 "Foundations of Ratio and Financial Analysis." [Paid] [Advanced]
• CFA Institute, Level I curriculum reading "Financial Analysis Techniques". The standard-setter's definitions of every ratio here, worth reading once for convention hygiene. [Paid] [Intermediate]

Free courses & videos:

• Aswath Damodaran's free "Accounting 101" class (YouTube playlist + session notes on pages.stern.nyu.edu). Sessions on ratio analysis and the financial-statements-to-analysis bridge. [Free] [Beginner]
• Zerodha Varsity, Module 3 "Fundamental Analysis," the financial-ratio chapters. Indian statements, Indian examples, gentle pace. [Free] [Beginner]
• Damodaran's "Ratings, Interest Coverage Ratios and Default Spreads" table (pages.stern.nyu.edu, updated each January; pull the live version). The synthetic-rating bridge you will use in M3.03. [Free] [Intermediate]

Ratings methodology (how professionals weight these exact ratios):

• S&P Global Ratings, Corporate Methodology and the ratios-and-adjustments supplement; Moody's sector rating methodologies. Free with registration; read one to see leases, pensions, and restricted cash adjusted by the people who do it for a living. [Free] [Advanced]
• CRISIL / ICRA / CARE rating rationales (free on their websites, searchable by company). Every Indian rationale discusses coverage, DSCR, and working-capital intensity of a real issuer in two pages. Read five. [Free] [Intermediate]

Filings & datasets to pull (the module's specimens):

• Avenue Supermarts AR and Costco 10-K (mini-project). HUL AR for trade-payables scale and the MSME note. Chipotle 10-K for the leases note behind WE1. InterGlobe Aviation (IndiGo) AR for the free vs restricted cash disclosure. Titan AR for inventory, gold-on-lease and customer advances. All [Free].
• screener.in (India) and stockanalysis.com (US) for ten-year ratio time series. Then recompute one year by hand and reconcile: vendors' conventions (leases in debt? revenue-based turns?) differ, and finding the difference is the exercise. [Free] [Intermediate]

Going deeper (optional, after the mastery check):

• Going deeper — Michael J. Mauboussin, writings on working capital, cash flow, and return on invested capital (Counterpoint Global / Morgan Stanley Investment Management, and his Columbia Business School notes; free PDFs, search the current titles) [Free] [Intermediate]. Connects the cash-conversion cycle and working-capital intensity to free cash flow and returns on capital, the conceptual bridge from this week's balance-sheet clock to the valuation work in Phase 3.
• Going deeper — Howard Schilit, Jeremy Perler & Yoni Engelhart, Financial Shenanigans (4th ed.), the cash-flow and working-capital chapters [Paid] [Intermediate]. How a stretched-payables "float," channel-stuffed receivables, and inventory games surface in exactly the ratios you just learned (rising DSO and DIO against sales); the trailer for M2.06–2.08.

The Modern Analyst's Addendum

Everything above teaches this skill from first principles, by hand; that is how you learn it, and the mastery check still tests it that way. This addendum shows how a working analyst amplifies the same skill today. It adds; it never replaces. (R1/R10)

AI-Augment this skill

{
  "skill": "Liquidity and coverage ratios, the five turnovers and their day-counts, and the cash conversion cycle read as a sentence about power in a value chain",
  "use": "Every ratio in this module is a convention fight before it is an arithmetic problem, and that is where a tool earns its place. (1) Before comparing two companies, ask it to list every convention choice embedded in a stated ratio: COGS or revenue in the inventory numerator, purchases or COGS for payables, averages or ending balances, 365 or 360 days, fiscal-year-end dates, and which choice each company's own filing or aggregator has made. Convention mismatch is the single commonest source of a wrong working-capital comparison, and it is entirely mechanical to enumerate. (2) Hand it a decomposed CCC move (the DIO, DSO and DPO contributions separately) and ask for every business explanation consistent with that specific pattern, then go find which one the filing supports; a CCC that improved purely on a DPO lurch has a short list of causes and §4.2's counterfeit float is on it. (3) Ask it to build the seasonality map for a peer set: fiscal year-ends, quarter-end balance-sheet dates, festival or holiday timing. The photograph problem is a calendar problem and calendars are what tools are for. (4) Ask it to draft the 'who finances whom' sentence for a value chain and then argue against its version.",
  "tools": ["Chat assistants: Claude, ChatGPT for convention enumeration and for candidate explanations of a specific CCC pattern", "Document/PDF AI: for the receivables and payables ageing schedules and the segment working-capital disclosures across a peer set", "Coding assistants: for the day-count panel and the cash-per-day conversion below"],
  "prompt": "Here are five years of DIO, DSO, DPO and CCC for <COMPANY> and three peers, with the formula and balance convention I used: <PASTE>. For the largest CCC move in the panel, list every operating or accounting explanation consistent with that specific pattern of component changes, ranked by how well each explains the SIZE of the move, and state for each what filing disclosure would confirm or eliminate it. Do not tell me which one is true, and do not use any figure that is not in my table.",
  "verify": "Recompute one company's full set from the filing by hand before trusting any panel; the day-counts are the easiest ratios in finance to get subtly wrong, and a single revenue-instead-of-COGS inventory numerator laundered 18 days of phantom speed in §4.5. Confirm that the payables convention is the same for every company in the comparison, that balance-sheet terms are averages, and that fiscal year-ends are close enough to compare at all. Then convert the days into money before you rank anything by severity, because a day is not a unit of cash (see the Quantitative lens). A model asked whether a CCC move is worrying will always have a view; only the ageing schedules and the segment disclosures can say (Guardrail 1-5).",
  "diy": "The gate is unaided: you compute all five turnovers with the correct numerators, build the CCC from a filing, and read the working-capital story off the balance sheet on paper. Nothing in this callout is admissible in that room."
}

Modern Data Analysis

By hand first. You walked one rupee through Kaveri's cycle, computed the five turnovers with the numerators defended, and built the DMart-versus-Costco table by hand. Keep doing that for the first company you meet in any sector. The code below is for the tenth, and for the ten years.

Today's workflow. Working-capital analysis is a panel problem disguised as a ratio problem. The level tells you the business model, and only the trend against peers tells you whether something is wrong. So the unit of work is a frame of company-years, and a single filing won't do. Two moves matter once the frame exists.

The first is conversion into cash, the step most analysts skip. One day of receivables is worth revenue ÷ 365, and one day of inventory is worth COGS ÷ 365. On Kaveri's figures a day of DSO is ₹20.00 crore and a day of DIO is ₹14.00 crore. So a 25-day CCC deterioration costs ₹500 crore if it's in receivables, but only ₹350 crore if it's in inventory.

The second is decomposing the move, which the CCC gives you for free because it's a sum. The change in CCC equals the change in DIO plus the change in DSO minus the change in DPO. So the panel can carry three attribution columns, and the warning in 4.6 about a CCC that "improved" on a DPO lurch becomes a filter you run every time. Add the growth-cash column, operating working capital as a percentage of revenue, at 6.8% for DMart against 0.67% for Costco. Now you can price the cash cost of a growth plan directly. DMart's next ₹10,000 crore of revenue traps about ₹680 crore before a rupee reaches shareholders.

Tools & sources (IN + US). pandas for the panel and the attribution columns, matplotlib for the day-count trend charts. For India, start with the balance sheet and Schedule III's mandatory trade-receivables and trade-payables ageing schedules. US filers don't give you that disclosure, and it's the best working-capital artefact in Indian reporting. Add the quarterly LODR results for the seasonal photographs, and the eleven mandated ratios with their >25%-change explanations. All of it comes from the annual report and the exchanges (BSE/NSE). For the US, use EDGAR XBRL companyfacts (Revenues, CostOfRevenue, InventoryNet, AccountsReceivableNetCurrent, AccountsPayableCurrent), the 10-K for the allowance and concentration disclosures, and the 10-Qs for the quarterly balance sheets a seasonal business demands. FRED (BUSINV, ISRATIO) gives the economy-wide inventory-to-sales context around any single company's DIO.

# Kaveri Consumer Products, the module's own figures, ₹ cr — R3: recomputed in-session
rev, cogs, inv, recv, pay, d_inv = 7300., 5110., 840., 600., 640., 80.
DIO, DSO = 365 * inv / cogs, 365 * recv / rev            # 60.0, 30.0
DPO = 365 * pay / (cogs + d_inv)                          # 45.0 on purchases (45.7 on COGS)
print(round(DIO, 1), round(DSO, 1), round(DPO, 1), round(DIO + DSO - DPO, 1))   # CCC 45.0
cash_per_day = {"DSO": rev / 365, "DIO": cogs / 365, "DPO": cogs / 365}
print({k: round(v, 2) for k, v in cash_per_day.items()})  # 20.0, 14.0, 14.0 ₹ cr per day
print(round(cash_per_day["DSO"] / cash_per_day["DIO"], 4))    # 1.4286 = 1/(1 - gross margin)
wc_pct = (inv + recv - pay) / rev                             # cash cost of growth
print(round(wc_pct * 100, 2), round(wc_pct * 10000))      # 10.96% -> ₹1,096 cr per ₹10,000 cr

Verify. Reconcile the panel's most recent year to a filing you spread by hand, and check the convention on every column. An aggregator that publishes revenue-based inventory turns will hand you 42 days where the correct figure is 60. You won't spot the error, because both numbers look plausible. Confirm the fiscal year-ends before you compare seasonal businesses at all. A US retailer closing end-January and an Indian company closing 31 March are photographs of different moments in their own cycles, so they can't be laid side by side. Where India's ageing schedules exist, use them to check DSO, and don't run the check the other way round. A stable DSO with a growing over-six-months bucket is the pattern the average hides, and the forensic work goes looking for it.

Quantitative lens

Two questions here can only be answered with numbers, and both change how you read a day-count.

A day isn't a unit of cash. One day of receivables ties up revenue ÷ 365, and one day of inventory ties up COGS ÷ 365. Their ratio is revenue ÷ COGS, which equals 1 / (1 − gross margin). At Kaveri's 30% gross margin a receivables day costs 1.43× what an inventory day costs. At a 50% gross margin it's 2.00×, at 70% it's 3.33×, and at a software company's 85% it's 6.67×. So ranking working-capital problems by days is systematically wrong, and the error grows with gross margin. A high-margin company with 20 days of inventory and 60 days of receivables has almost all of its cash trapped on the receivables side, even though the day-counts don't scream it. For a low-margin distributor inventory days and receivable days are nearly interchangeable in cash terms. Always convert to money before you rank severity.

How big is a CCC move that means nothing? Section 4.6 is right that trend beats level. That raises a question the level never asks: what does an ordinary year look like? Take Kaveri's 45-day cycle and let each of the three components wobble lognormally by 8% a year, independently, with no deterioration at all. Run it and the standard deviation of the year-on-year CCC change is 6.50 days. In 12.4% of clean years the cycle moves more than 10 days one way or the other, and the 95th percentile of the annual change is 11.0 days. So a 10-day CCC move is no tripwire. It's a Tuesday. The alarm case from earlier, an FMCG company drifting from 20 to 45 to 70, is a 7.7 standard-deviation move against that noise. That's what makes it a fire alarm, while a single 8-day wobble isn't one.

These limits are real, and they aren't small. The 8% component noise is an assumption. The right figure is the company's own historical dispersion over a period without a regime change. Measure it yourself and don't import it. The three components are also assumed independent, and that's wrong in the direction that matters. A demand shock moves inventory and receivables together, so real ΔCCC has a fatter tail than this simulation, and a large move is somewhat less surprising than 7.7 standard deviations suggests. Finally, none of this arithmetic can tell a deteriorating business from a deliberate strategy change. DMart chose its 29-day cycle. So the statistic sizes a move and never interprets it. It tells you when to open the ageing schedule, and a screen shouldn't claim more than that.

Do it in code. Build the day-count panel, convert every component to cash per day, decompose each year's CCC change into its three contributions, and standardise the changes against the company's own historical dispersion, not a rule of thumb. See G1/QM1.01 and G2/DA1.05.

Where this goes next: galaxy cross-links

• G2 · DA1.02 Pandas Data Wrangling — the multi-year, multi-company day-count panel with its attribution columns, and the ageing schedules as data rather than as pictures in a PDF.
• G2 · DA1.05 Statistics in Code — measuring a company's own working-capital dispersion so that "a big move" has a definition instead of a feeling.
• G1 · QM1.01 Probability, Distributions & Estimation — why a threshold on a noisy three-part sum fires as often as it does, and how to choose one that catches events rather than variance.
• G1 · FI1.02 Duration & Convexity — coverage and solvency from the lender's side of the table: 4.4 asks whether the flow carries the claims, and this is where the claims themselves are priced and their sensitivity measured.

Teach it back & journal

Feynman prompt. Write a one-page explainer for a smart 15-year-old titled "The shop that pays its bills after it has already sold your stuff." Use a lemonade stand, or a kirana shop. Explain why a shop that owes more than it owns right now (current ratio below 1) can be the safest shop on the street. Explain what DIO, DSO, and DPO are as three clocks. Explain why the shop's owner cares whether the CCC is +45 or −5 when she wants to open a second stall. No jargon allowed until the last paragraph, where you may introduce the terms and attach them to what you already explained.

Journal prompt. Take the business you know best. Your own consulting work qualifies perfectly. Map its cash conversion cycle: how many days from doing the work (inventory ≈ work-in-progress) to invoicing (DSO starts) to collection, against how long you take to pay your own vendors and salaries (DPO). Who finances whom between you, your clients, and your team? Then pick the single lever that would most change the cash the business needs to grow 30%. It might be faster billing or faster collection. It might be negotiating retainers (customer advances). Write down what that tells you about why companies you analyze fight so hard over payment terms.

This module's flashcards and mastery quiz are wired into the app: see the node's Quiz and Reviews.