Stress Testing: How Severe-but-Plausible Cases Are Designed
A stress test asks whether a business or a plan survives a severe but plausible event, not whether it stays profitable. The event is designed from the exposure map: break the thing most depended on, then trace the damage through to cash and obligations. The result is a survival answer, months of cover remaining, and it is only honest if the supports that would fail together are broken together.
The whole sequence converges here. Concentration named the thing most depended on; liquidity supplied the cash, the line and the collections clock; leverage supplied the obligations that never stop; scenarios taught case consistency; the exposure map supplied the forces that fail together. A stress test assembles all of it and asks one blunt question: does the business survive, and for how many months?
How is a stress test different from a bad-case scenario?
The question changes, and the arithmetic changes with it. Every household that has asked how many months it would last if the salary stopped tomorrow has run a stress test, and what nobody asks in that sentence is whether the year's savings target will be met. The scenario asks what remains in a likely world; the stress test asks whether survival is possible in an unlikely one, and the pass mark moves from profit to months of cover. A fire drill is not a weather forecast.
Tessora Weaves' downside scenario cut orders 30 per cent and profit before tax (PBT) survived at Rs 58,00,000. Is the stress test just a bigger version of the same case?
What does severe but plausible actually mean?
Two words, two tests, and both must pass. Severe enough to break the largest dependency: for Tessora Weaves, anything that leaves Meridian Retail Group's 60 per cent intact is not yet a stress. Plausible enough that a named cause could do it: a retailer exiting a supplier is an event that happens in the real world every season, so "Meridian exits" qualifies. An asteroid does not, and a 5 per cent dip does not either, from the other direction. Severity without plausibility is fantasy; plausibility without severity is a scenario wearing a hard hat.
Which of these qualifies as severe but plausible for Tessora Weaves: a 5 per cent revenue dip, Meridian Retail Group exiting entirely, or all global trade halting forever?
Where does the stress event come from, if not imagination?
From the exposure map, mechanically. The largest dependency is already known from the concentration analysis: one buyer at 60 per cent. The design rule has three steps. Break the thing the map shows the greatest dependence on, name a cause that could break it, and trace the damage line by line. No brainstorming session required, and that is the point: imagination produces stresses people can live with, while the map produces the one they actually fear.
Why is the pass mark survival rather than profit?
Because in the stress world, profit is already gone, and pretending otherwise wastes the exercise. The PBT line does not decide whether Tessora Weaves exists in eighteen months. Cash, the line and collections decide it, set against the obligations that do not stop: the fixed costs and interest keep arriving whether or not the revenue does, the lesson leverage taught in its cruellest form. So the unit of answer is months of cover: how long the supports last against the monthly shortfall. Months buy options, a new buyer found, costs cut, the loan renegotiated. The stress test measures how many options the business gets.
In the stress world, why measure months of cover instead of the year's profit or loss?
Why must supports that fail together be broken together?
The liquidity loop returns here: the event that removes the revenue is the same event that slows collections and makes the bank nervous about the line. So an honest stress breaks them together, and the difference is enormous. Tessora Weaves' stress, run twice, shows it. Naively, Meridian exits but everything else behaves: cash Rs 1,20,00,000, the undrawn line Rs 1,50,00,000 and the full Rs 8,00,00,000 of receivables all arrive as planned against a monthly shortfall of roughly Rs 79,00,000: about 13 months of cover. Honestly, the same exit slows collections, only Rs 1,60,00,000 of the book arrives promptly, and the line is frozen: 3 to 4 months. The ten-month gap between those two answers is the value of the joint-failure discipline, and it is the difference between a plan and an obituary.
Two numbers are in question here: months of cover at full severity with supports intact, and with supports jointly failed.
Choose the severity. Then break the supports honestly.
One input: the share of revenue lost. The toggle then runs the same severity with collections slowed and the line frozen, the joint failure an honest test requires.
How many months does Tessora Weaves survive its own worst dependency?
The full worked stress, in one table, both runs. The event: Meridian Retail Group exits, 60 per cent of revenue. Remaining contribution about Rs 80,00,000 a month against Rs 1,59,00,000 of obligations: a monthly shortfall of roughly Rs 79,00,000 that starts on day one and does not negotiate. The naive run survives about 13 months; the honest run 3 to 4. The honest answer is the planning number, and what it buys is the to-do list: a second anchor buyer before the stress, not after it; a committed line that survives bad news; a cost base that can shed weight fast.
| The stress, worked | Supports intact | Jointly failed |
|---|---|---|
| Monthly shortfall | Rs 79,00,000 | Rs 79,00,000 |
| Cash | Rs 1,20,00,000 | Rs 1,20,00,000 |
| Undrawn line | Rs 1,50,00,000 | frozen |
| Receivables arriving | Rs 8,00,00,000 | Rs 1,60,00,000 |
| Months of cover | about 13 | 3 to 4 |
The honest stress gives 3 to 4 months. Name the most valuable thing Tessora Weaves can do with that answer, today, before any stress arrives.
What does a household stress test look like?
Exactly the same machine, smaller numbers, and running it once for a single household is the fastest way to make the method permanent. The largest dependency in a one-salary household is the salary, so the stress is: it stops. The obligations that continue: the equated monthly instalment (EMI), rent, school fees, the kitchen. The supports: the emergency fund, anything borrowable, anything sellable fast. Naive months of cover: the fund divided by the monthly bill. The same downturn that takes the job also tightens every employer in the sector, makes the personal loan harder, and finds the gold price on its own schedule rather than the household's. So the supports have to be broken jointly here too. The honest number is smaller, and it is the one that should size the emergency fund.
Notice what the exercise changes: nothing about the risk, everything about the preparation. Preparation is stress testing's entire product, at every scale from a kitchen to a bank.
A household computes 8 months of cover from its emergency fund, assuming a new job in the same sector arrives whenever needed and a personal loan is always available. Name the kind of stress test this is.
The error that gets made, and what it costs
The stress test that passed because it was gentle: a 5 per cent revenue dip tested against intact collections and an open line, reported as stressed, and filed. The business then meets the real version, where the supports failed along with the revenue. It carries into that event a survival estimate three times too generous. The gentle test did not merely miss the risk; it certified its absence.
The cost is not the event. The cost is the ten months of preparation the gentle test said were unnecessary.
A stress report shows a pass. Name the two questions that decide whether the pass means anything.
Close the loop on the whole sequence: which earlier idea tells a stress designer what to break, and which one explains why the supports vanish together?
References
| Source | Document | Where |
|---|---|---|
| Reserve Bank of India (RBI) | Published financial stability material where stress testing is used | rbi.org.in |
Tessora Weaves Private Limited and Meridian Retail Group are invented.
Educational material. Not advice on any investment, tax, budget or market position.
