Risk Concentration

Anyone who invests money or runs a company can never fully avoid losses. What matters is whether possible losses are spread broadly or bundled at a single point. This bundling is exactly what is called risk concentration. One example: a supplier generates 80 percent of its revenue with a single car manufacturer. If it loses that customer, it isn’t just part of the business that disappears, but almost all of it. The more common everyday term for this is lump risk.

Why a single failure can bring down entire companies

The old rule “don’t put all your eggs in one basket” describes the underlying problem well. Anyone who spreads their money across thirty stocks can cope with the collapse of a single company. Anyone who puts everything into one stock is completely dependent on its fate. The expected return can be the same in both cases. The possible total loss is not.

For banks, this point is so important that it is regulated by law. Supervisory authorities limit how much money a bank may lend to a single borrower. In the EU, this large exposure limit is roughly a quarter of the bank’s equity capital. The idea behind it: even if a large borrower goes bankrupt, the bank should still remain standing. Without such limits, a single default could bring down an entire bank – and with it, its customers.

A common misconception is to assume risk concentration only occurs with investments. It appears just as much in supply chains, customers, locations, or software providers. Dependence on a single person, such as a founder who holds all the customer relationships, is also a form of it.

When risks only appear to be spread out

To measure risk concentration, one looks at what share of a total holding falls on individual positions. Analysts check, for example, what percentage of revenue the three largest customers make up. Or how much of a loan portfolio sits in a single industry. The higher this share, the stronger the concentration. Some companies disclose such figures in their annual reports.

The harder part is hidden concentrations. A portfolio with fifty stocks looks broadly diversified. But if it’s fifty technology stocks, they all depend on the same factors: interest rates, chip demand, regulation. Experts then speak of correlation, meaning that prices move in lockstep. In that case, the diversification exists only on paper.

Effective countermeasures work in three ways. One deliberately spreads more broadly, for example across several industries and regions. One sets fixed upper limits, for example a maximum of five percent of assets per single position. Or one hedges the risk, for example through a second supplier for an important component. This hedging costs money in good times – it only pays off in an emergency.

From the cloud to the chipmaker

In the tech industry, risk concentration is currently a constant topic. Advanced AI chips are almost entirely designed by one company and manufactured by a single contract manufacturer in Taiwan. A natural disaster or political conflict there would hit the entire global economy. That is precisely why the EU and the US are financing their own chip factories with billions in funding.

The situation is similar with cloud services, meaning rented computing power from external data centers. Three providers share the bulk of the market. If a major region of one provider fails, thousands of websites and apps are often offline at the same time. Such outages have occurred several times in recent years, each lasting a few hours.

In the news, you’ll typically encounter this term in three contexts: bank supervision and stress tests, warnings about dependencies in supply chains, and the analysis of stock indices. In the US S&P 500 index, a considerable share of the total value now falls on a handful of technology corporations. Anyone who buys this index is therefore diversifying less broadly than the number 500 might suggest.