HOW IT WORKS / 3 MIN READ

Five risks can fail as one.

Combining five exposures does not automatically spread the risk. If they share a failure trigger, losses can arrive together. This small model keeps each exposure’s failure chance the same while changing how often they move together.

Try the numbers ↓

Split the exposure into five equal parts.

Each part has the failure chance you choose and loses 60% of its value if it fails. The expected loss is exposure × failure chance × 60%.

Dial up shared failures.

At 0% linkage, the five failures are independent. At 100%, all five share a single yes-or-no failure event. Values between the two mix those models while preserving each exposure’s individual failure chance.

Look at the worst 1%.

CVaR at 99% is the average loss in the worst 1% of modeled outcomes. The illustrative premium adds 10% of that tail loss to the average loss. The chart shows every possible loss level and its probability.

THE SIMPLE MATH

Demo premium = average loss + 10% × average loss in the worst 1%

A worked example

For $1,000,000 exposure and a 5% failure chance, average loss is $30,000 at every linkage setting. At 100% linkage, all exposures fail together 5% of the time, losing $600,000. The worst-1% average is then $600,000 and the demo premium is $90,000.

What this example assumes

The six possible loss levels are calculated exactly, rather than sampled randomly. Independent losses use the binomial distribution; linkage mixes in a common shock. Severity is fixed at 60%, tail confidence at 99%, and tail loading at 10%. This is not a calibrated insurance quote; expenses, time horizon and an additional correlation charge are omitted.