Credit & Distressed Analysis

Phase 11 · Elective track · Node E11.01 · ~14 hours · Milestone: two credit workups (one ₹, one $) + a timed credit-committee drill

Where it sits. IL&FS group entities held AAA ratings within weeks of defaulting to D in September 2018, and the market's spread had been whispering long before the letters changed. Credit investors reprice survival before equity investors accept it.

This is an elective, but it is the elective that changes how you read every company for the rest of your career. Phases 1–3 taught you to spread statements, judge earnings quality, and value a business as an equity holder, the owner of the residual claim. Credit flips the chair around and seats you where the lender sits: senior to the equity, capped on the upside, and paid to obsess about the one question the equity story often waves away. Will this company be able to pay what it has promised, on the dates it promised, through a downturn it did not forecast?

You will learn to read a balance sheet as a stack of promises with a priority order, to price the risk in those promises as a spread, to test the covenants that police them, to trace what each claimant actually recovers when the promises break, and to reverse-engineer how Moody's and CRISIL turn all of this into a letter grade. The payoff is not a second career in fixed income (though this is its foundation). The payoff is that an equity analyst who can read credit sees the downside first: refinancing walls, covenant tripwires, priority subordination. Such an analyst never again mistakes a levered equity's fair-weather returns for a margin of safety.

Suggested hour budget. Core content (read actively, compute along): 5h · worked examples: 2h · practice set: 3h · applied mini-project (two credit workups + the committee drill): 3h · flashcards + mastery check: 1h.

App link. This module seeds the Credit & Solvency branch of the skill galaxy: the coverage/leverage ladder calculators, the covenant-headroom sandbox, the recovery-waterfall builder, and the timed credit-committee mode (defend a rating and a spread against seeded objections). The artifact it produces, artifacts/credit-memo-template.md, becomes a permanent companion to the M2.08 quality-of-earnings checklist: QoE tells you whether the numbers are true; this memo tells you whether, true or not, the company can pay.

A note on synthetic credits and real mechanics, before anything else. Every company worked below is synthetic, invented so that the arithmetic is clean and nobody is defamed. The mechanics are real: the Insolvency and Bankruptcy Code's Section 53 waterfall, the way Moody's and CRISIL build a rating from a business-risk and financial-risk scorecard, the difference between a maintenance and an incurrence covenant, the US absolute-priority rule. Each is described as accurately as a teaching text can. Rating symbols, statutory waterfall orders, and covenant conventions all drift over time and across jurisdictions; wherever the text states a live mechanic it flags it " ", and you should, against the current Code, the current rating scale, and the actual indenture in front of you. Numbers in the worked examples were each computed and checked in Python; reproduce them, because reproducing them is the skill.

Learning objectives

You will be able to:

1. Adopt the debt investor's lens. Articulate the payoff asymmetry of a fixed claim (capped upside, full downside), explain why that asymmetry makes creditors risk-first readers, and state where a lender's questions diverge from an equity holder's on the same set of statements.
2. Decompose a credit spread into its economic parts: risk-free rate + expected loss (PD × LGD) + a risk/liquidity premium. Compute a corporate bond's spread over the sovereign benchmark from its yield, apply the credit triangle (spread ≈ PD × LGD) in both directions, and estimate a bond's price move from a change in spread using spread duration.
3. Build and read the coverage and leverage ladders: Debt/EBITDA and Net Debt/EBITDA, EBITDA/Interest and (EBITDA − capex)/Interest, FFO/Debt and RCF/Debt. Define FFO and RCF cleanly, adjust leverage for capitalised leases under Ind AS 116 / IFRS 16, and know why FCF and cash conversion outrank any single accrual-based ratio.
4. Distinguish maintenance from incurrence covenants, read the core negative covenants of an indenture (negative pledge, restricted payments, debt incurrence, asset sales, change-of-control), compute covenant headroom and the EBITDA decline that trips a maintenance test, and explain why covenants are an equity issue (they transfer control before insolvency does).
5. Run a priority-of-claims waterfall for both an Indian IBC Section 53 liquidation and a US absolute-priority restructuring, compute the recovery percentage for each class of claimant, identify the fulcrum security where value breaks, and distinguish structural from contractual subordination.
6. Reconstruct a rating the way Moody's and CRISIL do: separate business risk from financial risk, apply a weighted scorecard to reach a standalone credit profile, then notch for parent or sovereign support. State too the known limits of ratings (lag, issuer-pays conflict, the AAA-to-D cliff) with the IL&FS precedent in mind.
7. Read the early signals of distress: spread and price thresholds, the maturity wall, covenant breaches, rating actions, the credit market repricing a name's paper before its equity accepts it. Describe the restructuring toolkit (amend-and-extend, debt-for-equity, DIP financing, cramdown) at a level that lets an equity analyst price refinancing risk before it becomes a headline.

Prerequisites & connections

Builds on. Phase 1's statements (the balance sheet is now a priority stack, not just assets = liabilities + equity; the cash-flow statement is now a debt-service test). Phase 2 entirely: M2.04 (CFO, FCF, cash conversion, the raw material of every coverage ratio), M2.05 (benchmarking, since spreads and leverage only mean something cross-sectionally), M2.06–M2.08 (quality of earnings; a lender who trusts an inflated EBITDA lends against a number that isn't there, and IL&FS, DHFL, Yes Bank, and Cox & Kings from the forensic capstone are credit failures first). Phase 3's valuation and cost of capital (the cost of debt you plug into a WACC is a credit spread; enterprise value is what the waterfall distributes). From Phase 1: M1.09 (consolidation, where structural subordination lives in the gap between standalone and consolidated debt) and M1.10 (the auditor's report, CARO, and the security and charge disclosures).

Feeds into. The sector playbooks (every sector has a characteristic capital structure and a characteristic way of dying: banks and NBFCs on ALM mismatch, infrastructure on refinancing, commodities on the cycle). The rapid-analysis teardown (a ten-minute solvency screen is this module compressed). Portfolio construction and risk (position sizing against downside, not just expected return). And directly into any distressed, special-situations, or high-yield mandate, for which this is the entry gate.

The one-sentence version of this module. Equity is a call option on the enterprise value struck at the total debt; what follows teaches you to value and police the debt, the thing that sets the strike, writes the rules, and takes the keys when the option expires worthless, so that you understand the company as its most senior, most skeptical, and most contractually powerful investors do.

Core teaching content

11.2 Spreads for credit: the price of default risk

A corporate bond's yield is not a single number; it is a stack, and pulling it apart is the first technical skill of credit. Conceptually:

Corporate yield = risk-free rate + credit spread, and credit spread ≈ expected loss + risk premium, where expected loss ≈ PD × LGD.

Take each layer. The risk-free rate is the yield on the sovereign benchmark of matching maturity and currency: the US Treasury for dollars, the Government of India (G-sec) for rupees. It is the time-value-of-money floor; every credit is priced as a spread over it. The credit spread is everything the market demands on top for taking corporate default risk: it is the number that moves when a company's fortunes change, and it is what a credit analyst is really trading in.

Decompose the spread itself. Probability of default (PD) is the chance the borrower fails to pay as promised over a horizon (usually annualised). Loss given default (LGD) is the fraction of exposure you lose if it does default, and crucially LGD = 1 − recovery rate, so a bond you expect to recover 40 cents on the dollar has an LGD of 60%. Their product, expected loss = PD × LGD, is the actuarially fair compensation: what you would charge if you were risk-neutral and could diversify away the idiosyncratic risk. But investors are not risk-neutral, and defaults cluster in recessions exactly when investors can least afford them, so the market charges more than expected loss: a risk premium (plus a liquidity premium for bonds that are hard to sell). That wedge, spread minus expected loss, is why credit held to maturity across a diversified book has historically earned a positive excess return. You are paid for bearing a risk you dislike, not merely for the losses you expect.

The credit triangle. The single most useful back-of-envelope relationship in credit is:

spread ≈ PD × LGD

It is an approximation (it ignores the risk premium, discounting subtleties, and assumes a flat hazard rate), but it is the analyst's mental slide rule, and it works in both directions. Given a market spread and an assumed recovery, you can imply the default probability the market is pricing: a bond trading at a 500 bp spread with an assumed 40% recovery (LGD 60%) is pricing an annual PD of roughly 5.00% ÷ 60% = 8.33%. Given your own estimate of PD and LGD, you can compute the spread you would demand and compare it to the market's. That comparison is the essence of relative-value credit. If you think PD is really 4% and LGD 60%, fair spread is 4% × 60% = 240 bp; a bond offered at 500 bp is cheap on your view, one at 150 bp is rich.

The spread ladder. Spreads are ordered by credit quality, and the ordering is the ratings scale made visible in price. In broad, illustrative terms (order-of-magnitude, and regime-dependent, so check against current market levels): in the US, investment-grade (IG) corporates might trade tens to low-hundreds of basis points over Treasuries, while high-yield (HY) trades in the high hundreds to thousands. The IG–HY boundary (BBB−/Baa3 to BB+/Ba1) is the single most important line in the market, because a fallen angel, a bond downgraded from IG to HY, is force-sold by mandate-constrained investors, which widens its spread mechanically. In India, corporate bonds are quoted as a spread over the G-sec of matching tenor, and the AAA → AA → A ladder widens at each step, with a pronounced cliff below A because much institutional money is rating-constrained. The shape and level of the whole curve moves with the cycle: spreads compress in booms (the "reach for yield") and blow out in stress (2008, March 2020, the 2018 IL&FS freeze in India's CP market).

Spread risk, not just default risk. You do not have to suffer a default to lose money in credit; you only have to suffer a spread widening. The sensitivity of a bond's price to a change in its spread is its spread duration: ΔPrice% ≈ −(spread duration) × (Δspread). A bond with spread duration 4.2 whose spread widens 150 bp loses roughly 4.2 × 1.50% = 6.3% of price, before any coupon. This is why credit desks track DTS (duration times spread) as their core risk unit: a low-spread bond with long duration and a high-spread bond with short duration can carry the same mark-to-market risk. For an equity analyst the lesson is adjacent but sharp. When a company's bond spreads widen, its cost of capital has risen and its refinancing just got more expensive, whatever the equity is doing, because the credit market is repricing survival probability in real time.

One practical note before the ladders. Every spread above is quoted off a yield, and a yield is computed from a settlement price that is not the price you see on a screen. The flat price a dealer quotes excludes accrued interest; the full price you actually pay includes it, and the accrual depends on which of three day-count conventions the instrument uses. FI1.05 §4.7 and §4.8 own that arithmetic, including why an Indian debenture on a 30/360 basis and a US Treasury on actual/actual give different accruals on the same dates. Get it wrong and every spread in this module is computed off the wrong number.

Two families of credit model, and why the module has quietly built both. Everything above is one of the two standard ways of modelling default, and it is worth naming both, because the profession sorts credit research into these families and expects you to know which one you are holding.

A structural model starts from the balance sheet and asks a question about the firm's assets. Robert Merton's 1974 formulation is the canonical one and it is exactly §11.8's sentence made computable: equity is a call option on the value of the firm struck at the face value of the debt, so default happens when the asset value at maturity falls below what is owed. Its inputs are the market value of the firm's assets, the volatility of those assets, the debt level and the horizon. Its outputs are an equity value, a default probability and a credit spread, all from one option-pricing formula. This module's addendum builds it and sweeps it: at a 30% asset volatility, moving leverage from 60% to 70% of firm value takes the implied spread from 40.1 bp to 129.4 bp, and holding leverage at 60% while moving asset volatility from 30% to 40% takes it from 40.1 bp to 173.7 bp. Read the two grids there rather than rebuilding them.

A reduced-form model refuses that question. It treats default as an unpredictable event arriving at a hazard rate, an instantaneous conditional intensity, and it calibrates that intensity to observed prices instead of deriving it from the balance sheet. Nothing in it says why the firm defaults. FI1.04 §4.9 builds this family properly, with the survival function, the risky annuity, the protection leg and a bootstrapped credit curve, and the credit triangle you just learned is its crudest special case: spread ≈ hazard × LGD is a one-period reduced-form model with a flat intensity.

Each fails in a characteristic way, and the failures are the reason both survive. The structural family's documented weakness is the credit spread puzzle: it underpredicts observed spreads, most severely for short-maturity investment-grade credits, because it prices only the expected-loss leg and a market spread also contains the risk and liquidity premia this section opened with. Eom, Helwege and Huang (2004) and Huang and Huang (2012) are the standard references; read them for magnitudes rather than quoting one from here. It also needs asset value and asset volatility, neither of which is observable, so both are backed out from equity prices under assumptions that do their own damage on a company with complex or off-balance-sheet debt.

The reduced-form family's weakness is the mirror image. It fits every price it is shown and explains none of them, so it cannot tell you that a credit is mispriced; if the market is wrong, the calibrated hazard rate is wrong in exactly the same direction. And a single spread quote cannot separate the hazard rate from the recovery assumption: FI1.04 shows a 210 bp five-year quote implying a hazard of 2.63% at 20% recovery, 3.50% at 40% and 5.25% at 60%. Any sentence of the form "the market is pricing an X% chance of default" is incomplete until the recovery assumption is stated out loud.

India changes the practical balance between the two. A reduced-form calibration wants single-name credit-default-swap quotes, and India's single-name protection market is thin, so the hazard rate has to be inferred from corporate bond spreads over the G-sec curve instead. Those spreads carry a liquidity premium that is larger and more variable than in the dollar market, which pushes the inferred hazard up and does so by an amount nobody can decompose from one quote. The structural route has its own local difficulty: promoter-group structures, pledged holdings and unlisted subsidiaries make "the market value of the firm's assets" a harder object than the formula assumes, which is E11.04's whole subject. Neither problem is a reason to skip the estimate. Both are reasons to state the assumption in the memo.

So use them for different jobs. Reach for the structural family when you want an opinion the market has not already given you, which is what an early-warning screen is; reach for the reduced-form family when you need a number consistent with tradeable prices, which is what marking a book or hedging a position requires. When the two disagree materially on the same credit, that disagreement is the trade idea, and the work is to decide which input you trust: your asset volatility, or the market's spread. The one thing neither family does is the thing §11.3 to §11.5 spend their time on. A model can price the default it has been told about. Only the ladders, the covenants and the waterfall tell you what happens to your claim when it arrives.

11.8 The synthesis: the equity analyst who reads credit first

Return to where §11.1 began: equity is the residual claim, a call option on enterprise value struck at the face value of the debt. Everything above is, for the equity analyst, a way of pricing that option honestly. The leverage and coverage ladders tell you how far the enterprise value can fall before the option goes out of the money. The covenants tell you that control can transfer to the lenders before the option formally expires, so the equity holder's fate is in the creditor's hands at the first maintenance breach, not the last. The waterfall tells you that when the option does expire worthless, the equity recovers essentially nothing, because the creditors take the whole enterprise. The rating and the spread tell you what the market's most senior, most skeptical investors think the survival probability is, a prior you ignore at your peril. And distress analysis tells you that the option can be destroyed not by a bad business but by a bad maturity schedule: a solvent company can die a liquidity death when the wall arrives and the market is shut.

The discipline this builds is a reading order. Before you value an equity, you now: (1) locate the debt, gross, net, and structurally senior; (2) build the coverage and leverage ladders, through the cycle, adjusted for leases and off-balance-sheet items; (3) find the maturity wall and ask who refinances it and at what spread; (4) read the covenants for the tripwires that transfer control; (5) sketch the waterfall to see what the equity recovers in the bad state. Only then do you price the good state.

The equity analyst who does this sees risk first, not because they are pessimists, but because they have learned, from the creditors, that the left tail is where fortunes are actually lost, and that the returns which do not compensate for it are a loan you are unknowingly making to someone more senior than you.

Common mistakes & how experts think differently

1. Reading EBITDA as cash. The novice levers off Debt/EBITDA and stops. The expert knows EBITDA is a pre-everything number (before interest, tax, working capital, and the capex the business cannot skip) and drops down the coverage ladder to (EBITDA − capex)/Interest, FFO/Debt, and FCF. A 4× EBITDA/interest that becomes 1.5× after maintenance capex is not comfortable; it is a warning wearing comfortable clothes.

2. Trusting net debt without asking where the cash is. Netting cash against debt assumes the cash is available to repay the debt. The expert asks whether it is trapped in a subsidiary, held overseas, pledged, restricted in a DSRA, or, worst, not real (the cash-yield test from M2.08 is a credit test). Net leverage computed on inaccessible cash is a number that lies in exactly the situations where it matters.

3. Ignoring structural subordination. The novice reads the consolidated balance sheet and assumes a holdco bond is as safe as the group's leverage implies. The expert computes standalone-vs-consolidated debt and knows that opco creditors are paid from opco assets first, so the holdco bond is structurally behind them, whatever its headline seniority. IL&FS holdco creditors learned this in the recovery, not before.

4. Confusing maintenance and incurrence covenants. Treating a high-yield bond's incurrence covenants as if they were maintenance covenants (or vice versa) mis-times the risk entirely. The expert knows a maintenance covenant trips on drift (an early-warning tripwire), while an incurrence covenant only tests on an action, so a cov-lite structure can let leverage climb for years with no tripwire at all, and the expert prices that missing alarm.

5. Mistaking a rating for an analysis. The novice outsources the judgment to AAA. The expert treats the rating as an informed prior with a known lag and a known conflict, and re-derives the credit, because IL&FS was AAA weeks before D, and the market's spread had been whispering long before the letters changed. A rating the market's pricing contradicts is a question, not an answer.

6. Pricing the mean and calling it a margin of safety. The equity habit is to model the base case. The credit habit, and the source of real margin of safety, is to underwrite the recession: trough EBITDA, a shut refinancing market, a covenant breach. The expert asks "does the capital structure survive the bad year?" not "how good is the good year?"

7. Reading leverage point-in-time on a cyclical. A commodity producer at the top of its cycle shows deceptively low leverage. The expert normalises to mid-cycle or trough earnings, because the debt is fixed and the EBITDA is not. Leverage is a through-the-cycle judgment, and the peak-cycle number is the one that gets credits downgraded eighteen months later.

8. Forgetting the maturity wall. Solvency (assets > liabilities) and liquidity (cash to pay what is due now) are different, and companies die of the second while solvent on the first. The novice checks leverage; the expert checks the maturity schedule and asks who rolls the paper coming due, at what spread, in what market: the IL&FS/DHFL failure mode.

9. Applying the credit triangle without respecting its assumptions. spread ≈ PD × LGD is a mental slide rule, not a pricing model. It ignores the risk and liquidity premia (which is why credit earns an excess return) and assumes a flat hazard rate. The expert uses it to sanity-check and to imply the market's PD, then remembers the residual wedge is real money and real risk, not noise.

10. Treating recovery as a constant. Assuming 40% recovery everywhere ignores that recovery is seniority-ordered, sector-specific, and procyclical, worse precisely when defaults cluster. The expert ties recovery to the collateral and the cycle, and knows that an asset-light company's "senior unsecured" can recover far less than the historical average when its value is a departing workforce.

11. Analysing one instrument in isolation. A bond does not default alone. Cross-default and cross-acceleration mean the whole capital structure moves together, and the fulcrum analysis requires the entire stack and an enterprise value. The expert maps every layer (first lien, second lien, senior unsecured, subordinated, structural) before pricing any one of them.

12. Thinking credit is only for bond investors. The costliest mistake for an equity analyst. The cost of debt in your WACC is a spread; enterprise value is what the waterfall distributes; the equity is the residual option struck at the debt. An equity analyst who cannot read the credit is pricing only the right tail of a two-tailed distribution — and the left tail is where the money is lost.

Worked examples

Six workups, each number computed and verified. Companies are synthetic; the mechanics are real. Three are in rupees (WE1, WE3, WE5) and three in dollars (WE2, WE4, WE6). Reproduce every figure: the reproduction is the skill.

Worked example 1: Bharat Cement (synthetic, ₹): the leverage & coverage ladder, mapped to a rating

You are a credit analyst covering Indian building materials. Bharat Cement Ltd (synthetic) reports (₹ crore): revenue 12,000; EBITDA 2,400; depreciation & amortisation 700; interest 520; cash taxes 300; total debt 8,000; cash & liquid investments 1,200; capex 1,500; dividends 150.

Step 1: the leverage rungs. EBIT = 2,400 − 700 = ₹1,700 crore; net debt = 8,000 − 1,200 = ₹6,800 crore.

• Debt/EBITDA = 8,000 ÷ 2,400 = 3.33×
• Net Debt/EBITDA = 6,800 ÷ 2,400 = 2.83×

Three-and-a-third turns gross, under three net: moderate for a capital-intensive but stable cash-generative business.

Step 2: the coverage rungs, descending.

• EBITDA/Interest = 2,400 ÷ 520 = 4.62×
• EBIT/Interest = 1,700 ÷ 520 = 3.27×
• (EBITDA − capex)/Interest = (2,400 − 1,500) ÷ 520 = 900 ÷ 520 = 1.73×

Watch the ladder fall: a comfortable-looking 4.62× headline collapses to 1.73× once the capex the plant genuinely needs is subtracted. This is the single most important number in the workup. Cement is capital-hungry, and fair-weather interest cover flatters it.

Step 3: the cash-flow ladder. FFO = EBITDA − interest − cash taxes = 2,400 − 520 − 300 = ₹1,580 crore.

• FFO/Debt = 1,580 ÷ 8,000 = 19.75%, mid-to-low investment-grade territory.
• RCF = FFO − dividends = 1,580 − 150 = ₹1,430 crore; RCF/Net Debt = 1,430 ÷ 6,800 = 21.03%.
• FCF = FFO − capex = 1,580 − 1,500 = ₹80 crore; FCF/Debt = 80 ÷ 8,000 = 1.00%.

Step 4: read it as a lender. Leverage is fine (2.8× net), FFO/Debt is adequate (~20%), but the credit's soft spot is laid bare by two numbers: (EBITDA − capex)/Interest = 1.73× and FCF = ₹80 crore. This company deleverages barely at all from internal cash after capex and dividends. It lives on continued market access and on the capex cycle turning down. Through-the-cycle, if EBITDA fell 20% in a demand downturn while capex stayed committed, FCF turns firmly negative and coverage approaches 1×.

Step 5: the indicative rating. Net leverage ~2.8×, FFO/Debt ~20%, headline coverage 4.6× but thin post-capex, on a business with decent scale but commodity cyclicality: this maps to a low investment-grade profile: call it BBB / CRISIL A-band standalone (indicative; a real rating runs the full scorecard of WE5 and the sector methodology). The credit memo's one-line verdict: "Adequate leverage and FFO/Debt for low-IG; the risk is not the balance sheet today but the thin post-capex coverage through a cement down-cycle — size and covenant the exposure to survive a 20% EBITDA drawdown."

Worked example 2: Meridian Fabricators (synthetic, US$): decomposing a bond's spread and the credit triangle

A Meridian Fabricators (synthetic) senior unsecured bond: face $1,000, annual coupon 7.5% ($75), 5 years to maturity, trading at a price of 97.0 ($970). The 5-year US Treasury yields 4.20%. Expected recovery on the bond in default is 40%.

Step 1: the yield. Solve for the yield y that prices the cash flows (75 for five years plus 1,000 at maturity) at 970. By numerical solution (bisection, verified in Python), YTM = 8.26%. (Sanity check via the approximation [C + (F−P)/n] ÷ [(F+P)/2] = [75 + 6] ÷ 985 = 8.22%, close, as expected.)

Step 2: the spread. Credit spread = YTM − risk-free = 8.26% − 4.20% = 4.06% = 406 bp. This 406 bp is what the market demands for Meridian's default risk over the Treasury: the number a credit analyst actually trades.

Step 3: the credit triangle, forward. With expected recovery 40%, LGD = 1 − 0.40 = 60%. The market-implied annual default probability is PD ≈ spread ÷ LGD = 4.06% ÷ 60% = 6.76%. So at these prices the market is pricing roughly a 6.8% annual chance of default on Meridian. (Check: PD × LGD = 6.76% × 60% = 4.06% = the spread, as constructed; the triangle is internally consistent.)

Step 4: the credit triangle, reverse (relative value). Suppose your fundamental work says Meridian's true annual PD is only 4.5% (its order book is more secure than the market fears). Your fair spread = 4.5% × 60% = 2.70% = 270 bp. The bond trades at 406 bp, which is 136 bp cheap to your view: you are being paid materially more than your estimate of expected loss, plus the risk premium. That gap is the thesis. (Were your PD estimate higher than implied, the bond would be rich and you would avoid or short it.)

Step 5: spread risk. The bond has spread duration ≈ 4.2. If the credit environment deteriorates and Meridian's spread widens 150 bp (to ~556 bp), the mark-to-market price move is ΔP% ≈ −(spread duration) × Δspread = −4.2 × 1.50% = −6.30%, a 6.3% capital loss with no default at all. The lesson for the equity analyst watching from across the capital structure: a 150 bp spread widening means Meridian's marginal cost of debt just rose by that much, its next refinancing is more expensive, and the credit market has revised survival probability upward. That is information the equity should not ignore.

Worked example 3: Girnar Steel (synthetic, ₹): the IBC Section 53 waterfall

A Girnar Steel Ltd (synthetic) has gone into liquidation under the IBC. The liquidator realises ₹8,000 crore from the asset sale. Claims (₹ crore): insolvency & liquidation costs 200; workmen's dues (24 months) 150; secured creditors who relinquished their security to the pool 6,000; other employee dues (12 months) 120; unsecured financial creditors 3,000; government dues 400.

Step 1: apply the Section 53 order, top down. Start with ₹8,000 crore.

• Tier 1 costs (₹200 crore): paid in full off the top. Remaining = 8,000 − 200 = ₹7,800 crore.
• Tier 2 pari passu workmen (150) + relinquishing secured (6,000) = ₹6,150 crore: available (7,800) exceeds the claim, so both are paid in full. Remaining = 7,800 − 6,150 = ₹1,650 crore.
• Tier 3 employee dues (₹120 crore): paid in full. Remaining = 1,650 − 120 = ₹1,530 crore.
• Tier 4 unsecured financial creditors (₹3,000 crore): only ₹1,530 crore is left. They receive 1,530 ÷ 3,000 = 51.0% pro rata. Remaining = ₹0.
• Tier 5 government dues (₹400 crore): nothing left → 0%.
• Equity: 0%.

Step 2: recovery table.

Step 3: read it. The fulcrum is the unsecured financial creditors at 51%, the class where value runs out. Everyone senior is whole; everyone junior (government, equity) is wiped. Two structural lessons the numbers teach: (1) a secured creditor who relinquishes to the pool sits in the high-priority Tier 2, so security is worth real recovery; (2) government dues rank below unsecured financial creditors under the IBC, a deliberate policy choice that reshapes the textbook stack. In a going-concern resolution (rather than this liquidation), the committee of financial creditors could approve a plan distributing value differently, subject to the Code's minimum-recovery floors for dissenting and operational creditors. Verify against the current Code.

Worked example 4: Cobalt Corp (synthetic, US$): covenant headroom and the EBITDA cushion

A Cobalt Corp (synthetic) term loan carries two maintenance covenants, tested quarterly: maximum total net leverage 5.0× and minimum interest coverage 2.5×. Current figures ($m): EBITDA 400; net debt 1,700; cash interest 120. A separate high-yield bond has an incurrence covenant: additional debt only if pro-forma net leverage stays at or below 6.0×.

Step 1: current covenant positions.

• Net leverage = 1,700 ÷ 400 = 4.25× vs the 5.0× cap → currently compliant, with 0.75 turns of room.
• Interest coverage = 400 ÷ 120 = 3.33× vs the 2.5× floor → compliant.

Step 2: the leverage cushion (how far can EBITDA fall?). The leverage covenant breaches when net debt ÷ EBITDA = 5.0, i.e. when EBITDA falls to 1,700 ÷ 5.0 = $340m. The tolerable decline = (400 − 340) ÷ 400 = 15.0%. A 15% EBITDA drop trips the leverage covenant.

Step 3: the coverage cushion. The coverage covenant breaches when EBITDA ÷ 120 = 2.5, i.e. at EBITDA = 2.5 × 120 = $300m. Tolerable decline = (400 − 300) ÷ 400 = 25.0%. A 25% drop trips the coverage covenant.

Step 4: which binds? The leverage covenant is tighter: it trips at a 15% EBITDA decline versus 25% for coverage. So the binding early-warning tripwire is net leverage, and the credit analyst monitors the leverage headroom, not the (looser) coverage headroom, as the leading indicator. An equity holder should note: a mere 15% EBITDA dip, an ordinary bad year rather than a catastrophe, hands control to the lenders.

Step 5: the incurrence contrast. The bond's incurrence covenant permits new debt up to a pro-forma 6.0× net leverage: max net debt = 6.0 × 400 = $2,400m, so new-debt capacity = 2,400 − 1,700 = $700m. But that is only tested when Cobalt tries to borrow. If instead EBITDA simply drifts down, the incurrence covenant is never tested at all, even as leverage climbs past 6×. That is the structural difference from the maintenance covenant, which trips on the drift itself. This is precisely why cov-lite structures removed the market's early-warning system.

Worked example 5: Deccan Alloys (synthetic, ₹): a rating scorecard, built and notched

Build an indicative rating for Deccan Alloys Ltd (synthetic), a mid-scale specialty-metals producer, using a simplified weighted scorecard. Map each factor to a broad-rating numeric where AAA = 1, AA = 3, A = 6, BBB = 9, BB = 12, B = 15 (lower is safer). Factor scores from the analysis: Scale = 9 (BBB, mid-size); Business profile = 6 (A, differentiated products, decent diversification); Leverage = 12 (BB, elevated at ~4×); Coverage = 9 (BBB); Cash flow / FFO-Debt = 9 (BBB).

Step 1: the weights (sector-calibrated; they sum to 1.00). Scale 0.15; Business profile 0.25; Leverage 0.25; Coverage 0.15; Cash flow 0.20.

Step 2: the weighted score.

Weighted score = 1.35 + 1.50 + 3.00 + 1.35 + 1.80 = 9.00.

Step 3: map to a standalone rating. A score of 9.00 falls in the BBB band (9–11) → indicated standalone credit profile = BBB. Note what drove it: the poor leverage score (12, BB) is exactly offset by the strong business profile (6, A), a textbook illustration of the two halves of a rating trading off. A balance sheet that alone would score BB is pulled up to BBB by a defensible franchise; equally, a great balance sheet cannot rescue a terrible business.

Step 4: notch for support. Deccan Alloys is 74%-owned by a large, investment-grade industrial parent with a record of supporting group companies and an explicit (though not guaranteed) statement of strategic importance. Under a support framework this justifies +1 notch of uplift from the standalone profile → final indicative rating BBB+ (Moody's equivalent Baa2 → Baa1). Had the parent instead been weaker than the subsidiary, the rating could have been capped below the standalone level. Support notching cuts both ways, and a CE (credit-enhanced) rating would apply only if a specific structure (guarantee/escrow) backed the instrument, labelled as such. Verify the current methodology's notching bands.

Worked example 6: Odyssey Media (synthetic, US$): the fulcrum security in distress

A Odyssey Media (synthetic) is being restructured. Its capital structure ($m): first-lien secured debt 600; second-lien secured debt 400; senior unsecured notes 500, for total debt of $1,500m. The restructuring turns on the estimated enterprise value (EV) of the reorganised business, which is contested. Distribute the EV down the absolute-priority waterfall for three EV scenarios: $1,650m (bull), $1,200m (base), $850m (bear).

Step 1: base case, EV = $1,200m.

• First lien (claim 600): paid in full → 100%. Remaining = 1,200 − 600 = 600.
• Second lien (claim 400): paid in full → 100%. Remaining = 600 − 400 = 200.
• Senior unsecured (claim 500): receives the remaining 200 ÷ 500 = 40%. Remaining = 0.
• Equity: 0%.
• Fulcrum = senior unsecured (40%). This class converts its notes into the equity of the reorganised Odyssey, capturing the recovery and the upside. Loan-to-own investors accumulate here.

Step 2: bear case, EV = $850m.

• First lien: 600 → 100%. Remaining = 250.
• Second lien (claim 400): receives 250 ÷ 400 = 62.5%. Remaining = 0.
• Senior unsecured: 0%. Equity: 0%.
• Fulcrum has migrated UP to the second lien (62.5%). As value falls, the fulcrum climbs the stack: the senior unsecured, safe at $1,200m, is wiped at $850m.

Step 3: bull case, EV = $1,650m.

• First lien 600, second lien 400, senior unsecured 500, all paid in full ($1,500m total). Remaining = 1,650 − 1,500 = 150 → the equity recovers $150m (the option is back in the money).

Step 4: recovery matrix.

Step 5: the investing read. The entire distressed thesis is an enterprise-value view expressed through the right layer. Bullish on EV? The senior unsecured notes, trading at (say) 40 cents, are a leveraged call on the recovery, so buy the fulcrum. Bearish? The senior unsecured is a zero and even the second lien is impaired, so stay senior or stay out. Note the equity analyst's punchline: across the base and bear cases the equity recovers nothing. The equity is a call option that is out of the money unless EV exceeds the $1,500m total debt, which crystallises the §11.8 lesson that equity is a residual struck at the debt.

Practice set

Fourteen problems, ramped: guided (P1–P5, method prompted), independent (P6–P10, you choose the tools), timed (P11–P14, ~6 minutes each, committee conditions). All companies synthetic. Work each fully before reading the solution; every number is verified.

P1 (guided). Konkan Cement (synthetic) reports EBITDA ₹1,800 crore, total debt ₹5,400 crore, cash ₹600 crore, and interest ₹360 crore. (a) Debt/EBITDA. (b) Net Debt/EBITDA. (c) EBITDA/Interest. (d) One sentence on where this credit sits.

Solution. (a) 5,400 ÷ 1,800 = 3.00×. (b) net debt = 5,400 − 600 = 4,800; 4,800 ÷ 1,800 = 2.67×. (c) 1,800 ÷ 360 = 5.00×. (d) Three turns gross, under three net, 5× interest cover: a solid mid-investment-grade profile on these headline metrics; the next questions are post-capex coverage and the maturity schedule.

P2 (guided). Konkan Cement adds: cash taxes ₹210 crore, dividends ₹90 crore (EBITDA ₹1,800, interest ₹360, debt ₹5,400 as above). (a) FFO. (b) FFO/Debt. (c) RCF and RCF/Debt. (d) Interpret.

Solution. (a) FFO = 1,800 − 360 − 210 = ₹1,230 crore. (b) 1,230 ÷ 5,400 = 22.78%. (c) RCF = 1,230 − 90 = ₹1,140 crore; RCF/Debt = 1,140 ÷ 5,400 = 21.11%. (d) FFO/Debt ~23% is comfortably mid-IG, and RCF/Debt barely falls after a modest dividend (~21%). This credit retains most of its internal cash for deleveraging, a creditor-friendly financial policy.

P3 (guided). A bond trades at a spread of 300 bp; expected recovery is 45% (LGD 55%). (a) Imply the market's annual PD via the credit triangle. (b) If your fundamental work says PD is really 8%, what fair spread do you demand, and is the bond cheap or rich? (c) Name one assumption the triangle ignores.

Solution. (a) PD ≈ spread ÷ LGD = 3.00% ÷ 55% = 5.45%. (b) fair spread = PD × LGD = 8% × 55% = 4.40% = 440 bp; the market offers only 300 bp, so on your (worse) view the bond is rich: you are being under-paid for the default risk you see; avoid or short. (c) It ignores the risk/liquidity premium (spread exceeds pure expected loss in reality) and assumes a flat hazard rate / no discounting subtleties.

P4 (guided). A corporate bond yields 9.1%; the matching-tenor G-sec yields 7.1%. The bond's spread duration is 6.5. (a) Compute the credit spread. (b) If the spread widens 80 bp, estimate the price change. (c) Did the company default?

Solution. (a) spread = 9.1% − 7.1% = 2.0% = 200 bp. (b) ΔP% ≈ −6.5 × 0.80% = −5.2%. (c) No. A 5.2% mark-to-market loss occurred with no default at all, purely from spread widening; spread risk is distinct from default risk, and for an equity analyst it flags a rising cost of debt and refinancing pressure.

P5 (guided). Vindhya Chemicals (synthetic) has a maintenance covenant: maximum gross Debt/EBITDA of 4.5×. Current gross debt ₹900 crore, EBITDA ₹240 crore. (a) Current leverage. (b) The EBITDA level that trips the covenant. (c) The tolerable EBITDA decline (%).

Solution. (a) 900 ÷ 240 = 3.75×, compliant. (b) breach at Debt ÷ EBITDA = 4.5, i.e. EBITDA = 900 ÷ 4.5 = ₹200 crore. (c) decline = (240 − 200) ÷ 240 = 16.67%. A 16.7% EBITDA drop hands the lenders control: an ordinary bad year, not a disaster.

P6 (independent). Aravalli Textiles (synthetic) liquidates under the IBC realising ₹5,000 crore. Claims (₹cr): costs 150; workmen (24m) 100; secured-relinquished 4,000; employees (12m) 90; unsecured financial 1,500; government 200. Compute each class's recovery and name the fulcrum.

Solution. After costs: 5,000 − 150 = 4,850. Tier 2 (workmen 100 + secured 4,000 = 4,100): available 4,850 > 4,100 → both 100%; remaining 4,850 − 4,100 = 750. Employees (90): 100%; remaining 660. Unsecured financial (1,500): receives 660 ÷ 1,500 = 44.0%, the fulcrum. Government (200): 0%. Equity: 0%. Lesson: relinquishing secured creditors sit in Tier 2 and recover fully; unsecured financial creditors are the fulcrum at 44%; government and equity are wiped.

P7 (independent). Nilgiri Retail (synthetic) reports EBITDA ₹500 crore (post Ind AS 116), borrowings ₹1,200 crore, and lease liabilities ₹800 crore. (a) Leverage on borrowings only. (b) Leverage including lease liabilities (as an agency would). (c) The understatement, in turns, and why it matters.

Solution. (a) 1,200 ÷ 500 = 2.40×. (b) (1,200 + 800) ÷ 500 = 2,000 ÷ 500 = 4.00×. (c) The borrowings-only view understates leverage by 4.00 − 2.40 = 1.6 turns. It matters because lease obligations are fixed, senior, contractual claims that must be serviced exactly like debt. A retailer or airline analysed on borrowings alone looks investment-grade while being, in economic substance, leveraged; the leverage is hidden in the lease-liability note.

P8 (independent). Score a credit with weights: business profile 0.30, leverage 0.30, coverage 0.20, cash flow 0.20; and scores (AAA=1…BB=12…): business 6, leverage 9, coverage 6, cash flow 12. (a) Weighted score. (b) Indicated band. (c) If liquidity is weak (a near-term maturity wall with thin committed lines), apply a one-notch adjustment and state the final rating.

Solution. (a) 6×0.30 + 9×0.30 + 6×0.20 + 12×0.20 = 1.80 + 2.70 + 1.20 + 2.40 = 8.10. (b) 8.10 sits at the A/BBB boundary (A band = 6–8), so indicated A− (low A, negative bias). (c) Weak liquidity is a standard override; −1 notch → BBB+. The weak cash-flow score (12) and the maturity-wall liquidity concern are what drag an otherwise A-ish business franchise down to crossover.

P9 (independent). Kaveri Power (synthetic): EBITDA ₹600 crore; capex ₹380 crore; interest ₹150 crore; cash taxes ₹70 crore; dividends ₹40 crore. (a) EBITDA/Interest. (b) (EBITDA − capex)/Interest. (c) FCF before dividends, and after. (d) The credit read.

Solution. (a) 600 ÷ 150 = 4.00×. (b) (600 − 380) ÷ 150 = 220 ÷ 150 = 1.47×. (c) FCF pre-dividend = 600 − 380 − 150 − 70 = ₹0; after dividends = 0 − 40 = −₹40 crore. (d) Headline 4.0× interest cover looks safe, but the business generates zero free cash after its capital programme and taxes, and turns cash-negative after the dividend. It deleverages not at all and depends on continued market access; the ₹40 crore shortfall must be funded with new debt or cash drawdown. Coverage without free cash flow is a mirage.

P10 (independent). Sutlej Retail (synthetic) has first-lien debt $500m and senior unsecured notes $700m (total $1,200m). Restructuring EV is contested. (a) Recoveries and fulcrum at EV = $900m. (b) At EV = $400m. (c) What migrates, and the general rule.

Solution. (a) EV 900: first lien 500 → 100%; remaining 400; unsecured 400 ÷ 700 = 57.1% → fulcrum = senior unsecured; equity 0. (b) EV 400: first lien 400 ÷ 500 = 80% → the fulcrum has migrated up to the first lien; unsecured 0; equity 0. (c) As enterprise value falls, the fulcrum security climbs the priority stack: the class that owns the reorganised equity moves senior. The distressed thesis is always an EV view expressed through whichever layer is the fulcrum at that EV.

P11 (timed, ~6 min). Chambal Industries (synthetic): net debt ₹3,200 crore; gross debt ₹3,600 crore; EBITDA ₹800 crore; interest ₹220 crore; FFO ₹480 crore. (a) Net Debt/EBITDA. (b) EBITDA/Interest. (c) FFO/Debt. (d) IG or HY?

Solution. (a) 3,200 ÷ 800 = 4.00×. (b) 800 ÷ 220 = 3.64×. (c) 480 ÷ 3,600 = 13.33%. (d) Four turns of net leverage, sub-4× coverage, and FFO/Debt at ~13% (below the ~15–20% IG floor) put this firmly in speculative-grade / high-yield (BB) territory. The cash-flow ratio confirms what the leverage turn hints at.

P12 (timed, ~6 min). A bond yields 8.6%; the risk-free benchmark yields 6.9%; your estimate of expected loss (PD × LGD) is 90 bp. (a) The spread. (b) The portion of spread that is not expected loss. (c) What that residual compensates.

Solution. (a) spread = 8.6% − 6.9% = 1.70% = 170 bp. (b) residual = 170 − 90 = 80 bp. (c) The 80 bp is the risk premium plus liquidity premium: compensation for bearing default risk that clusters in bad times and for holding a less-liquid instrument, over and above the actuarial expected loss. This residual is why diversified, held-to-maturity credit earns a positive excess return.

P13 (timed, ~6 min). Godavari Infra (synthetic): Loan A has a minimum DSCR covenant of 1.25×; Loan B cross-defaults to Loan A. Cash available for debt service ₹250 crore; total debt service ₹210 crore. Loan A outstanding ₹1,000 crore; Loan B ₹1,400 crore. (a) DSCR. (b) Breach? (c) Consequence via cross-default, in ₹.

Solution. (a) DSCR = 250 ÷ 210 = 1.19×. (b) 1.19× < 1.25× → breach of Loan A. (c) The cross-default clause means the Loan A default triggers a default on Loan B as well; both may be accelerated, so the company faces potential acceleration of 1,000 + 1,400 = ₹2,400 crore at once. A single covenant miss can bring the whole stack due. This is why covenants are mapped across the entire debt structure, not instrument by instrument.

P14 (timed, ~6 min). Marina Shipping (synthetic): a senior unsecured claim of $800m recovers $320m in the restructuring; a CDS contract references it on a $500m notional. (a) Recovery %. (b) LGD. (c) CDS payout to the protection buyer.

Solution. (a) 320 ÷ 800 = 40.0%. (b) LGD = 1 − 0.40 = 60.0%. (c) A single-name CDS pays (1 − recovery) × notional = 0.60 × 500 = $300m to the protection buyer on the credit event (cash-settled at the recovery determined by auction). The recovery rate simultaneously sets the bondholder's loss and the CDS payout: LGD is the hinge between the two.

P15 (guided, ~8 min: revenue-bond coverage and the two covenants). (numeric; tolerance ±0.005 on each coverage ratio and ±0.05 percentage points on each cushion.) A synthetic municipal water and sewer system has gross revenues of $148,000,000 and operations and maintenance costs of $92,500,000. Senior-lien annual debt service is $34,800,000 and subordinate-lien annual debt service is $8,600,000. The senior lien carries a 1.20× rate covenant; the all-in test is 1.10×. (a) Net revenues available for debt service. (b) Senior and all-in coverage. (c) The percentage fall in net revenues that breaches each covenant. (d) Which cushion you quote in the memo, and why.

Solution. (a) 148,000,000 − 92,500,000 = $55,500,000. Note the order: operations and maintenance sits above debt service in the flow of funds, which is the opposite of a corporate structure and the reason a muni analyst starts from net rather than gross. (b) Senior 55,500,000 ÷ 34,800,000 = 1.594828×; all-in 55,500,000 ÷ 43,400,000 = 1.278802×. (c) The senior covenant binds at net revenues of 1.20 × 34,800,000 = 41,760,000, a fall of 24.7568%. The all-in covenant binds at 1.10 × 43,400,000 = 47,740,000, a fall of 13.9820%. (d) Quote 13.9820%, the tighter of the two. The senior number flatters the credit by ignoring a lien that is contractually junior to you but economically ahead of the system's ability to keep operating, and a subordinate-lien breach is a real event with real consequences even when your own coupon is safe.

P16 (independent, ~8 min: the rate covenant working in your favour, and the additional-bonds test). (numeric; tolerance ±0.005 on the coverage ratio, ±0.05 percentage points on the rate rise, ±$50,000 on the dollar amounts.) Same system. A drought year takes gross revenues to $132,000,000 with operations and maintenance unchanged. (a) Stressed net revenues and stressed senior coverage. (b) The revenue increase the 1.20× rate covenant obliges the issuer to produce, and that increase as a percentage of stressed gross revenue. (c) Separately, at full revenues, how much new senior annual debt service a 1.25× additional-bonds test would permit. (d) What (b) tells you that a corporate covenant never would.

Solution. (a) Net 132,000,000 − 92,500,000 = $39,500,000; senior coverage 39,500,000 ÷ 34,800,000 = 1.135057×, below the 1.20× covenant. (b) The covenant requires net revenues of $41,760,000, so the shortfall is 41,760,000 − 39,500,000 = $2,260,000, which is 1.7121% of stressed gross revenue. (c) 55,500,000 ÷ 1.25 = $44,400,000 of total permitted senior debt service, so headroom for new senior debt service is 44,400,000 − 34,800,000 = $9,600,000 a year. (d) A corporate maintenance covenant is a tripwire: breach it and control passes to the lenders (§11.4). A rate covenant is an obligation to act: breach it and the issuer must raise prices, which for an essential monopoly service it can usually do. The credit question therefore moves from "can the company survive the breach" to "can the issuer legally and politically deliver a 1.71% tariff increase," and on that question a regulated issuer and an unregulated one are different credits at identical coverage.

P17 (independent, ~10 min: scoring a sovereign, and computing the fiscal input). (numeric; tolerance ±0.01 on the scorecard, ±0.05 percentage points on the debt-dynamics figures.) A sovereign scores 3 on economic strength, 4 on institutions and governance, 5 on fiscal strength and 4 on susceptibility to event risk. The scale runs 1 to 6 with 1 strongest, and the four groups are equally weighted. Its debt is 82.0000% of gross domestic product and the nominal effective interest rate on that debt is 7.2000%. Nominal growth is 9.8000% and the primary balance is a deficit of 1.5000% of output. (a) The weighted scorecard score, and which factor actually binds. (b) The change in debt to output this year and the debt-stabilising primary balance. (c) Redo (b) with the effective rate at 12.5000% after a confidence shock, growth unchanged. (d) State what the pair of answers means for the rating.

Solution. (a) 0.25(3 + 4 + 5 + 4) = 4.000000. The binding factor is fiscal strength at 5, because the event-risk and fiscal groups act as constraints rather than as averages; a committee that produced a 4.0 weighted score and then rated the sovereign at the level implied by its fiscal-strength score alone would be behaving normally. (b) The automatic term is (0.072 − 0.098) ÷ 1.098 × 0.82 = −1.941712%, so debt falls −1.941712 − (−1.500000) = −0.441712 points to 81.558288%, and the debt-stabilising primary balance is a deficit of 1.941712%. Growth above the interest rate is carrying the whole credit. (c) The automatic term flips to (0.125 − 0.098) ÷ 1.098 × 0.82 = +2.016393%, debt rises 3.516393 points, and stabilising it now requires a primary surplus of 2.016393%, a swing of 3.516393 points of output. (d) Nothing about the country's economy changed between (b) and (c); only the rate it borrows at did. So a rating that reads the current trajectory as strong is reading a g > r regime that the market can end unilaterally, and the correct memo line is that fiscal strength here is conditional on market access rather than on the budget. That conditionality is exactly what "susceptibility to event risk" is for, and it is why it caps rather than averages.

P18 (independent, ~8 min: an Altman score, and the question it does not answer). (numeric; tolerance ±0.005 on Z, ±0.5 bp on the loss figures.) A synthetic listed Indian manufacturer has total assets ₹4,250 crore and total liabilities ₹2,890 crore. Its working capital is ₹180 crore, retained earnings ₹615 crore and earnings before interest and tax ₹388 crore. Market capitalisation is ₹2,460 crore and sales are ₹5,180 crore. (a) Compute Z on the original public-manufacturer coefficients and place it against the 1.81 and 2.99 cuts. (b) Recompute Z after a 30% fall in the share price alone, and state what share of the original Z came from the market term. (c) Its bonds trade at a 410 bp spread; you assign the score band a one-year default probability of 4.5000% and a loss given default of 55.00%. Compute expected loss and the residual. (d) Say whether the bond is attractive, and what the score contributed to that answer.

Solution. (a) The terms are 1.2(180/4,250) = 0.050824, 1.4(615/4,250) = 0.202588, 3.3(388/4,250) = 0.301271, 0.6(2,460/2,890) = 0.510727 and 1.0(5,180/4,250) = 1.218824, so Z = 2.284233, in the grey zone. (b) Market capitalisation falls to ₹1,722 crore, the fourth term to 0.357509, and Z = 2.131015, a fall of 0.153218 with no change to any financial statement. The market term was 22.3588% of the original score, so roughly a fifth of a "fundamental" score is a share price. (c) Expected loss = 0.045000 × 0.550000 = 2.475000% = 247.5000 bp; the residual is 410.0000 − 247.5000 = 162.5000 bp, and expected loss is 60.3659% of the spread. (d) The score does not answer the question. It ranks this credit as weaker than a 2.99-plus name, and that is all it does. Whether the bond is attractive depends on whether 162.5000 bp is adequate compensation for the risk and liquidity premium you are being asked to bear, which is a comparison against other bonds with similar expected loss (§11.2). A default model ranks survival; only the spread against your own expected loss ranks return, and the weakest credit in a set is regularly the best investment in it.

Applied mini-project: the credit workup and the committee drill

Budget ~3 hours. You will produce two credit workups (one Indian rupee credit, one US-dollar credit) and defend one of them in a timed credit-committee drill. The point is to make the reading order of §11.8 automatic and to produce a credit memo a portfolio manager could act on.

Companies. Use real, publicly-listed issuers with traded debt (pull the last two annual reports / 10-Ks, the debt maturity schedule from the notes, the latest rating rationales from CRISIL/ICRA/CARE or Moody's/S&P where free, and current bond/CDS levels where available). These are teaching specimens, not recommendations. Suggested candidate pools, so pick one from each:

• India (₹): Tata Steel (cyclical, lease and pension adjustments, through-the-cycle leverage), Bharti Airtel (capex intensity, spectrum liabilities, coverage ladder), Shriram Finance or an NBFC (ALM ladder, funding mix; use the financial-sector lens, not Altman), Grasim/UltraTech (capital structure across a group), or a REIT (lease-backed, DSCR-driven).
• US ($): Ford or GM (finance-arm consolidation, cyclical auto leverage), a BB-rated high-yield issuer of your choice (incurrence covenants, refinancing wall), Kraft Heinz (fallen-angel history, brand-value vs leverage), Netflix (content-asset capitalisation feeding into FFO), or a utility (regulated cash flow, high leverage, strong coverage).

Deliverables (per company):

1. The debt map: every layer of the capital structure (first lien / second lien / senior unsecured / subordinated / structural), gross and net debt, and the maturity schedule with the wall identified. Note the standalone-vs-consolidated gap where relevant.
2. The coverage & leverage ladders: Debt/EBITDA, Net Debt/EBITDA, EBITDA/Interest, (EBITDA − capex)/Interest, FFO/Debt, RCF/Debt, FCF/Debt, all computed, lease-adjusted, and shown through the cycle (a normalised or trough-EBITDA row, not just the latest year).
3. The covenant read: the key financial and negative covenants (maintenance vs incurrence), current headroom on the binding covenant, and the EBITDA decline that would breach it.
4. The spread & rating: the current spread over the sovereign benchmark, the credit-triangle-implied PD, the current agency rating(s), and your own indicative scorecard rating with any notching, plus a one-line note on whether the market's spread agrees with the rating.
5. The downside / recovery sketch: a priority-waterfall recovery estimate for the senior instrument in a stress scenario, and the identity of the fulcrum.
6. The one-page credit memo, conclusion first: your rating and spread view, the single biggest risk (usually the maturity wall or a binding covenant), what would change your mind, and, for the equity crossover, one sentence on what the credit picture implies for the equity holder.

Scoring rubric (apply per workup; 2 points each, 12 total; pass ≥ 9/12 with no criterion at 0):

The committee drill (~30 min). Present one workup as if to a credit committee and defend it against three seeded objections: (i) "your EBITDA is peak-cycle — what does the trough do to coverage and the covenant?"; (ii) "the rating is investment-grade but the spread says crossover — who is right and why?"; (iii) "walk the recovery if EV comes in 30% below your base — where's the fulcrum then?" Score yourself on whether each answer is quantified (a number, not an adjective) and reproducible.

Case Lab: run case CL-11.1 in the app's Cases tab for this module.

Reading & resources

• Moody's and S&P Global Ratings: published sector rating methodologies (free on the agencies' sites; registration). These are the actual scorecards: factor weights, sub-factor bands, notching frameworks. Read one corporate methodology end-to-end (e.g., a manufacturing or building-materials methodology) and reproduce its grid on a company you know. The single best way to understand how a rating is really made. [Free] [Intermediate]
• CRISIL / ICRA / CARE rating rationales and criteria (crisil.com, icra.in, careedge.in; free). India-specific: read a handful of rationales across the AAA→BB spectrum, plus the criteria documents on CE ratings, group support, and the financial-sector (NBFC/bank) approach. Grade each rationale against this module's ladders. [Free] [Intermediate]
• The Insolvency and Bankruptcy Code, 2016: Section 53 (waterfall) and the CIRP provisions (ibbi.gov.in, free). Read the statutory waterfall directly, then read one or two IBBI-published resolution outcomes to see recovery percentages in practice. Verify the current text; it has been amended repeatedly. [Free] [Advanced]
• Moody's / S&P annual default and recovery studies (free summaries). The empirical backbone: default rates by rating, recovery rates by seniority and cycle, rating-transition matrices. This is where your PD and LGD priors should come from — not from memory. [Free] [Advanced]
• Frank Fabozzi (ed.), The Handbook of Fixed Income Securities. The standard reference for yield, spread, spread duration, and the mechanics of the instruments. Use it as a lookup, not a cover-to-cover read. [Paid] [Intermediate]
• Moyer, Distressed Debt Analysis: Strategies for Speculative Investors. The practitioner's text on the fulcrum security, valuation in restructuring, and the negotiation dynamics: the deepest treatment of §11.7 in one place. [Paid] [Advanced]
• Moody's, Frydman & others / the Merton framework (any corporate-finance text's structural-model chapter). Equity as a call option on enterprise value struck at the debt (the §11.8 idea, formalised). Read it once to understand why credit and equity are two claims on one distribution. [Free] [Advanced]
• Marti Subrahmanyam / Edward Altman lectures and the Altman Z-score papers (free). For the quantitative-screen complement; recall from M2.07 that Z-scores are a solvency screen and are skipped for financials. Use the ALM/divergence panel there instead. [Free] [Intermediate]
• A live indenture and a live credit agreement (SEC EDGAR for US; company/BSE filings for India, free). Pull one high-yield bond indenture and one leveraged-loan agreement and read the covenant sections. Nothing teaches maintenance-vs-incurrence like the actual contract language. [Free] [Advanced]
• CFA Institute curriculum, Fixed Income: credit analysis and credit strategies readings. A clean, exam-grade synthesis of spread decomposition, the credit triangle, and rating logic, aligned to this module's framework. [Paid] [Intermediate]

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": "The equity-versus-credit dual read of one company: the credit spread's PD×LGD decomposition, the coverage/leverage ladder, covenant headroom, the IBC/absolute-priority waterfall and the fulcrum security, rating-scorecard reconstruction, and equity as a call option on the firm struck at the debt",
  "use": "This module's real AI use is document extraction, not analysis. (1) Pull the maturity schedule, the maintenance-vs-incurrence covenant triggers, and the negative-covenant baskets (restricted payments, debt incurrence, asset sales) out of a filed indenture or credit agreement exhibit — genuinely slow manual work that a tool can turn into a table you then check line by line against the exhibit's own section numbers. (2) Build the coverage/leverage input sheet from a 10-K or annual report — the line items with note references, not the ratios — the same 'input sheet, not the score' discipline this corpus already applies to forensic screens: you compute FFO/Debt yourself so one mis-pulled receivable can't move a number you never looked at. (3) Translate a CRISIL/ICRA/Moody's rating rationale into the business-risk/financial-risk factor language of §11.6 so you can attempt to reproduce its scorecard, then check the reproduction against the agency's own published methodology weights, not against the tool's say-so. Never ask it to judge whether a spread is 'right,' to run the waterfall, or to price the Merton model in its head — that arithmetic is exactly what this module makes you own.",
  "tools": ["Document/PDF AI — extracting maturity schedules and covenant triggers from a filed indenture or credit agreement exhibit, with a section or page reference per row", "Chat assistants — Claude, ChatGPT for the coverage/leverage input sheet and for translating a rating rationale into scorecard factor language", "Coding assistants — for the Merton engine, the covenant-breach probability calculation and the waterfall/fulcrum solver below"],
  "prompt": "From the attached [indenture / credit agreement exhibit], extract ONLY: the maturity schedule (instrument, principal, maturity date), every maintenance and incurrence covenant with its exact numeric trigger, the negative-covenant baskets (restricted payments, debt incurrence, asset sales), and the change-of-control terms — one row per item with a section or page reference. Do NOT compute any ratio, any headroom, or any rating. Where a term is not stated, write NOT STATED.",
  "verify": "Recompute every ratio, headroom figure, and PD yourself from the extracted inputs — this module's whole discipline is that the mechanics are exact and checkable, so an unchecked tool output defeats the point of learning them. Confirm every extracted covenant trigger against the actual exhibit page, because a maintenance-vs-incurrence misclassification changes which early-warning tripwire fires (§11.4) — the single costliest extraction error available. And cross-check any AI-summarized rating rationale against the agency's own published methodology document before trusting a reproduced scorecard weight; a plausible-sounding weight that is not the agency's own produces a rating that looks reasoned and is fabricated.",
  "diy": "The gate is unaided: you build the full coverage and leverage ladder from raw statements, distinguish maintenance from incurrence covenants and compute headroom by hand, run the IBC Section 53 or absolute-priority waterfall to find the fulcrum, reconstruct a rating from a weighted scorecard and notch it for support, and state and apply the credit triangle from memory. Nothing in this callout is admissible in that room."
}

Modern Data Analysis

By hand first. You built the coverage and leverage ladder rung by rung, derived the credit triangle and used it in both directions, ran the IBC Section 53 cascade to find the fulcrum, computed covenant headroom as an exact percentage decline, and built a rating from a weighted scorecard. One idea in the module, though, was only ever stated in words: §11.8's "equity is a call option on enterprise value struck at the face value of the debt." Code is what turns that sentence into a price.

Today's workflow. The Merton model prices exactly that option. Give it the same three things that already drive every rung of your ladder (a firm value, a debt level, and a volatility) and it returns an equity value, a risk-neutral default probability, and an implied credit spread, from one option-pricing formula. That is the structural complement to the credit triangle's market-implied read of §11.2: the triangle backs a PD out of an observed spread; Merton derives a spread from the balance sheet and a volatility estimate, with the market's mood nowhere in the inputs. Layer the same machinery onto §11.4's covenant headroom: WE4 computes the exact EBITDA decline that trips a covenant, which is arithmetic, not probability. Whether that decline is likely needs a volatility assumption for EBITDA, an input the headroom calculation never has to state, because "how far can EBITDA fall before the covenant breaks" is answered with one number, not a distribution. Code makes the missing assumption explicit and impossible to smuggle in silently.

Tools & sources (IN + US). numpy/scipy.stats.norm for the Merton solver and the covenant-breach probability, scipy.optimize.brentq for YTM and implied-volatility solves, pandas for the input sheet; India: CRISIL, ICRA and CARE's published rating rationales (crisil.com, icra.in, careedge.in, unusually detailed and free, per this module's own Reading & Resources) for the business-risk/financial-risk scorecard language, the IBBI's quarterly newsletter (ibbi.gov.in) for realised CIRP outcomes: actual recovery percentages and time-to-resolution by sector, the honest counterweight to any assumed LGD, RBI's Financial Stability Report for aggregate SMA (special-mention-account) and large-exposure trends across the banking system, and CCIL/FBIL for the G-sec and AAA-corporate curves that anchor the risk-free leg of every spread; US: EDGAR full-text search for indentures and credit agreements filed as exhibits (10-K/8-K Exhibits 4.x and 10.x, the actual contract language, not a summary), Moody's and S&P's published sector rating methodologies (the real scorecards, weights included), and FRED's BAMLC0A0CM (US corporate OAS) and BAMLH0A0HYM2 (US high-yield OAS) alongside DGS10 for the risk-free leg every spread in this module is measured against.

# Merton structural model: equity as a call on firm value struck at the debt — R3, in-session
import numpy as np
from scipy.stats import norm

def merton(V0, D, sigma_V, r, T):
    d1 = (np.log(V0/D) + (r + 0.5*sigma_V**2)*T) / (sigma_V*np.sqrt(T))
    d2 = d1 - sigma_V*np.sqrt(T)
    E0 = V0*norm.cdf(d1) - D*np.exp(-r*T)*norm.cdf(d2)      # equity = call value
    B0 = V0 - E0                                             # risky debt = firm - call
    pd_rn = norm.cdf(-d2)                                    # risk-neutral P(V_T < D)
    spread = -np.log(B0/D)/T - r                             # yield on B0 minus the risk-free leg
    return E0, B0, pd_rn, spread

# a synthetic 1-year credit: firm value 10,000, asset vol 30%, Rf 7% -- leverage swept 40%-90%
for lev in (0.40, 0.50, 0.60, 0.70, 0.80, 0.90):
    E0, B0, pd, sp = merton(10000, lev*10000, 0.30, 0.07, 1.0)
    print(f"leverage {lev:.0%}  PD {pd*100:5.2f}%  spread {sp*10000:6.1f}bp")

Verify. Prove the engine against the module's own worked figures before trusting it on a real credit. WE2's Meridian Fabricators bond reproduces exactly: YTM 8.26%, spread 406bp, credit-triangle PD 6.76%, and a −6.30% price move from a +150bp widening at spread duration 4.2. WE4's Cobalt Corp reproduces exactly: the leverage covenant trips at a 15.0% EBITDA decline and the coverage covenant at 25.0%, the anchor for the breach-probability extension below. WE5's Deccan Alloys scorecard reproduces exactly: weighted score 9.00, mapping to the BBB band. All three worked examples check out before the Merton engine is trusted on a company that matters.

Quantitative lens

The spread is not linear in either leverage or volatility, and the Merton engine puts a number on how not-linear. Hold asset volatility at 30% and raise leverage from 60% to 70% of firm value, ten points, and the implied spread does not rise gently: it goes from 40.1bp to 129.4bp, a 3.23× increase. Hold leverage at 60% and raise asset volatility from 30% to 40% instead, and spread goes from 40.1bp to 173.7bp, a 4.33× increase.

Neither move is a stress scenario; both are within an ordinary cyclical swing for a levered industrial name. And both are properties of the same option-pricing formula that makes §11.8's metaphor precise: a credit spread is levered twice, once by how much debt sits on the balance sheet and once by how uncertain the asset value backing it is.

Equity's payoff is right-skewed; debt's is left-skewed, and simulation puts an exact number on the shape, not just the word "asymmetric." Simulate two million draws of firm value at the same one-year horizon (asset drift 9%, other Merton inputs unchanged) and split the payoff the way §11.1 describes in prose: equity gets max(V_T − D, 0), debt gets min(V_T, D). Equity's payoff skewness is +1.01; debt's is −9.30, a far more extreme, one-sided shape, because debt's distribution is overwhelmingly a single narrow high-probability outcome (paid in full) with a long thin catastrophic tail on default, while equity's is spread smoothly across the whole upper half of the firm-value distribution once the option is in the money. The simulated physical default probability, 3.19%, matches the closed-form distance-to-default calculation of 3.196% to within 0.01 points of Monte-Carlo noise on two million draws, confirmation the simulation is specified correctly, not a number that merely looks right. Decompose the unconditional expected loss on the debt (₹20.38 cr) into frequency times severity and the credit lens sharpens further: it is a 3.19% probability times ₹638.85 cr of loss conditional on default, a conditional severity 31× the unconditional figure (exactly 1/PD, by construction). Reading "expected loss" as one small smooth number is reading the frequency and discarding the severity; the credit lens reads the severity first, because that is the number that arrives exactly when the cushion is gone and refinancing is shut.

WE4's covenant headroom is exact arithmetic; the probability of needing it is not, and the gap is worth showing once. The module computes, correctly and by hand, that Cobalt Corp's leverage covenant trips at a 15.0% EBITDA decline and its coverage covenant at 25.0%. Neither number says how likely a 15% or 25% decline is. That needs a volatility for EBITDA, an input the headroom calculation never uses and the module never supplies, because it doesn't have to. Assume EBITDA is lognormal around its current level with zero expected drift (the through-the-cycle convention §11.3 already argues for) and the breach probability is a one-line lognormal calculation. At an assumed 18% annual EBITDA volatility, the leverage covenant breaches with probability 20.8% and the coverage covenant with 6.6%. Move the assumption to 30% volatility (still inside the range a genuinely cyclical business could show) and the same fixed 15%/25% headroom now implies breach probabilities of 34.8% and 20.9%, a 1.67× and 3.18× rise respectively on an unchanged covenant. The exact same headroom, and the probability of needing it moves by that much on a volatility nobody observes directly. That is the honest content of "covenant headroom under a stress": the headroom is arithmetic; the likelihood of needing it is an estimate, and the estimate dominates the answer.

Honest limits. The Merton model prices default at a single horizon against a single zero-coupon-equivalent debt level, assumes constant asset volatility in a frictionless market with no jumps, no strategic default, and no refinancing option. None of that is true of Cobalt Corp, Meridian Fabricators, or any real borrower with a maturity schedule instead of one bullet payment. Its best-documented empirical failure is systematic: structural models of this family are well known to underpredict observed market spreads, especially for short-maturity, investment-grade credits, the "credit spread puzzle" catalogued by Eom, Helwege & Huang (2004, Review of Financial Studies) and examined further by Huang & Huang (2012, Review of Asset Pricing Studies); consult the papers directly for their current point estimates, because this addendum states the documented direction of the finding, not a magnitude it has verified itself. The gap is usually attributed to the risk and liquidity premia §11.2 already told you sit inside a market spread on top of expected loss, and Merton, by construction, prices only the expected-loss leg. And the recovery and default statistics that discipline any of this, LGD by seniority and the IBBI's realised CIRP recoveries, are cohort- and vintage-dependent everywhere, and the Indian series is short: the Code itself dates to 2016, so any Indian recovery statistic is drawn from far fewer full credit cycles than the Moody's/S&P long-run averages §11.5 already flags as "verify vintage and market." Treat every PD and every breach probability above as a property of the stated assumptions, not a forecast of any company's survival.

Do it in code: price equity as a call, read off the risk-neutral PD and the implied spread, and cross-check both against the credit triangle's market-implied read; turn any covenant's fixed percentage headroom into a breach probability by attaching a volatility assumption and stating it out loud; and decompose expected loss into frequency times severity before calling a number small. See G1/DV1.02, G1/QM1.01 and G2/QD2.02.

Where this goes next: galaxy cross-links

• G1 · DV1.02 Options Pricing & Greeks, the Black-Scholes machinery this addendum borrows wholesale: §11.8's "equity is a call option" stops being a metaphor once you can price it.
• G1 · QM1.01 Probability, Distributions & Estimation: the lognormal and normal-distribution machinery behind distance-to-default and the covenant-breach probability above, derived rather than assumed.
• G1 · FI1.02 Duration & Convexity: the same convexity logic that makes a bond's price nonlinear in yield is what makes a credit spread nonlinear in leverage and volatility; the bond-price side of this addendum's option-price side.
• G2 · QD2.02 Risk Systems, VaR & Expected Shortfall: the portfolio-level generalisation of the single-name loss distribution this addendum builds one credit at a time.
• G2 · CS2.06 Software Engineering & AI-Assisted Development: turning the Merton and covenant-breach engines into tested, reusable functions rather than one-off notebook cells, the discipline that makes a credit desk's tools survive turnover.

Teach it back & journal

Feynman prompt. Write a one-page explainer for a smart 15-year-old titled "Who gets paid first when a company runs out of money — and why the owners get paid last." You must use three ideas from this module in plain words: the priority stack (why a secured lender stands ahead of an owner), the fulcrum (the point in the line where the money runs out, and how the person standing there ends up owning the company), and the maturity wall (how a company that is not even broke can still fail because a big loan comes due on a day nobody will lend). No jargon: if you cannot explain why "equity is last in line" without saying "residual claim," you do not own it yet. End with the five-step reading order from §11.8 in words a teenager could follow.

Journal reflection. Open your decision journal and write a credit-lens entry: (1) pick a company you currently like as an equity, then locate its debt, its net and structural leverage, and its nearest maturity wall, and write one sentence on whether the equity return you expect is partly compensation for a refinancing risk you had not priced; (2) name one ratio from your past equity work that you now realise you read as an optimist (EBITDA as cash? peak-cycle leverage? net debt on trapped cash?) and how the credit lens changes it; (3) record your current mental model of one live rating you rely on. Do you treat it as analysis or as a prior with a lag and a conflict? Date it, and set a reminder to reread this entry the next time you find yourself admiring a levered equity's returns.

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