Concentration Risk: How Exposure Clusters and How It Is Measured

Concentration risk is the exposure that builds when too much of a business or a portfolio depends on one thing: one customer, one supplier, one lender, one market. Concentration is measured by the share of the largest exposure and by adding up squared shares, so a few large exposures score far higher than many small ones. The danger is not that the large exposure is weak, but that everything moves at once when it moves.

A value can move. When many rupees of value depend on the same thing, they stop being many separate uncertainties and become one large one. Counting names hides this; measuring shares reveals it. Two standard measures do the work, each has a place where it goes blind, and a lender or an analyst sitting across the table reads both before the conversation starts.

What is concentration risk, and what counts as an exposure?

A familiar case makes the point. A household where one person earns the entire salary is carrying concentration risk: every equated monthly instalment (EMI), every school fee, every plan in that house leans on one employer's decision. Nobody in that house needs a formula to feel it. Finance simply gives the feeling a name and a number.

The name is exposureAny single point that a material share of money, revenue or value flows through. If that point fails or changes behaviour, the value moves with it.: any single point that a material share of value flows through. For a business: a customer, a supplier, a lender, a geography, a product line. For an investor: a holdingOne investment inside a portfolio: one company's shares, one bond, one property., a sector, a currency. Concentration risk is what builds when one of those points carries a share large enough that its behaviour becomes the behaviour of the whole, the way that one salary is the behaviour of the whole household.

Tessora Weaves, an invented garment exporter, makes the idea concrete. The firm earns Rs 48,00,00,000 a year from exactly three buyers, and one of them, Meridian Retail Group, is 60 per cent of everything. The widths in the figure below carry the whole of the idea.

Why does clustering matter more than the size of any one exposure?

Picture a tea stall that has run for years outside one office building. Best location in the area, loyal crowd, steady cash. Then the office shifts to the other side of the city, and the stall's entire market walks away in a single month. Was the office a bad customer? No, it was a wonderful customer. The stall did not have a customer problem. The stall had a clustering problem.

The same logic at Tessora Weaves' scale: if Meridian Retail Group changes its buying programme, Rs 28,80,00,000 of revenue moves in the same season, for the same reason, in the same direction. The concern is never that the big exposure is weak. The big exposure may be the strongest name on the book. The concern is the size of the swing one decision can produce, and clustering is what makes the swing large.

Notice what stays behind when the swing arrives. Revenue is the moving part; fixed costsCosts that stay the same whatever revenue does: rent, salaries, the interest bill. Fixed costs are due in full even in a bad season. are not. Tessora Weaves carries Rs 18,00,00,000 of fixed costs a year, and a lost customer does not take those with it. Moving revenue set against immovable costs is why a concentrated book is a survival question and not just a growth question.

How is concentration measured with the largest share, and what does it miss?

The simplest measure is read straight off the mix: the share of the largest exposure. Tessora Weaves' largest-customer share is 60 per cent. The test needs no formula. The largest share answers the first question a lender or an analyst actually asks: how much of this business walks out of the door if the biggest name does?

Its blind spot is everything after the first name. Watch the two books below: one is split 40/40/20, the other is 40 with ten small buyers of 6 per cent each. The largest-share test scores them identically at 40. Yet the first book can lose its top two names, and 80 per cent of revenue, in one season. The second book cannot lose more than 46 from any two. A measure that sees one name cannot tell these books apart, and the second measure exists for exactly that reason.

How does the squared-shares index work, and why squares?

Square each exposure's percentage share, then add the squares. The result runs from near zero, everything spread thin, to 10,000, everything on one name. The measure is the Herfindahl-Hirschman Index (HHI), the standard way to compress a whole book into one concentration number.

The squares are the point, so sit with them for a moment. A 60 per cent share contributes 3,600 while a 30 per cent share contributes 900: doubling one share quadruples its contribution. The damage a single exposure can do rises faster than its size, so large exposures are punished more than proportionally. The punishment is what the risk deserves. The same arithmetic run in reverse produces the curve below, and its shape is worth memorising: splitting into equal exposures helps enormously at first and almost not at all after five.

What does Tessora Weaves' customer mix score, and what does the score mean?

Work it through once by hand; every number here can be redone from the table. Squares of 60, 25 and 15 are 3,600, 625 and 225. The sum is 4,450 of a possible 10,000. Ten equal customers would score 1,000; a single customer would score 10,000. So measured, not counted, Tessora Weaves sits closer to a one-customer business than to a spread one, and the revenue at riskThe rupees of revenue that leave if one named exposure leaves. The simplest way to convert a share into money. behind the largest share is Rs 28,80,00,000. The revenue behind one name, not the count of three, is the risk statement.

Where do businesses cluster without noticing?

A founder can name the customers from memory, so customers are only the first place to look. The same clustering builds quietly on every other side of the business. One mill supplying 70 per cent of the yarn is a concentration in the cost line. One bank providing the only working capital lineA borrowing limit a bank gives a business for day-to-day needs. The bank can freeze or cut it, and usually can at exactly the wrong moment. is a concentration in survival itself. All revenue landing in two export regions is a concentration in geography; one product category is a concentration in demand.

Each of these is measured exactly the same way: shares of the relevant total, largest share first, squares where one number is wanted. The discipline is to run the measure on every side of the business, not only the side that appears on the sales report.

Where does concentration hide behind many names?

Think of ten shops inside one mall. Ten owners, ten signboards, ten separate businesses on paper. Now the mall loses its anchor store and footfall halves. How many of those ten shops have a bad quarter? All ten, together, for one reason. The shops never shared a customer; they shared a driver.

The measures above count names, and names can lie. A shared driver is exactly what they cannot see. Ten customers who all sell to the same end shopper, ten holdings that all move on the same rate decision, ten lenders funded from the same market: each is one exposure wearing ten labels. Here is the trap, and it is worth sitting with: a book can score a beautiful 1,000 on HHI while everything on it responds to a single underlying force.

Tessora Weaves is the live example. Its three buyers are three names in three countries, but Meridian Retail Group, Nordhaven Stores and Calluna Home all sell to the same end marketThe final buyers a product ultimately reaches. A business can have many direct customers who all depend on the same end market.: US and EU retail shoppers. In a season when that demand falls, all three cut orders together. Measured by names the index is 4,450. Measured by driver it is close to 10,000. Reading exposure to a shared driver is set out under factor exposure. For now the rule is: measure the shares, then ask what sits underneath them.

How do lenders and analysts actually read these numbers?

A lender reads the largest-customer share as a repayment question. Some loan agreements carry a covenantA promise written into a loan agreement. Every payment so far may have been on time. Breaking the covenant still lets the lender demand repayment early or reprice the loan. that caps it: let the top buyer pass an agreed share of revenue and the loan can be repriced or recalled, because the bank knows it is effectively lending to Meridian Retail Group's purchasing department without being able to see it.

An analyst reads the same numbers as a valuation question. Companies disclose heavy customer dependence in their filings. The earnings of a business earning Rs 48,00,00,000 with one buyer at 60 per cent can vanish in one decision, and the business is therefore worth less than the same rupees spread across ten. The number does not need to be bad to matter; it needs to be priced. Concentration is not a verdict but a measurement, and the people on the other side of the table are already making it.

Tessora Weaves Private Limited, Meridian Retail Group, Nordhaven Stores and Calluna Home are invented.
Educational material. Not advice on any investment, tax, budget or market position.