Overview: Premium pooling creates the financial structure; the law of large numbers provides the mathematical engine that makes that structure stable. This section explains the principle and why it is so central to the insurance business model.

The law of large numbers is a theorem in probability that states: as the number of independent, similar observations (or trials) increases, the average of those observations will converge toward the true expected value. For insurance purposes, this means that as the number of similar policyholders in a pool grows, the insurer's actual loss experience will more closely match its predicted loss experience.

Here is the core intuition: for any individual policyholder, loss is essentially binary and unpredictable. A 45-year-old nonsmoker either dies this year (loss: the face amount) or does not (loss: zero). Nobody can predict which. But across 500,000 such policyholders, the statistical fluctuation around the average death rate becomes tiny. The insurer does not need to know who will die; it only needs to know how many will die on average, and history provides a reliable answer.

The mathematical key is how variance behaves when you average many independent outcomes. For a single policyholder, the variance (a measure of unpredictability) in loss is very large relative to the expected loss. But the variance of the average loss per policyholder shrinks as you add more people to the pool — it falls in proportion to 1 divided by the number of policyholders. Double the pool, halve the relative uncertainty. With hundreds of thousands of policyholders, the average loss becomes highly predictable, and the insurer can set premiums with confidence.

A Worked Example

Suppose an insurer offers auto collision coverage. Based on historical data, 2% of similar drivers will have a covered accident in a given year, with an average claim cost of $8,000. The expected loss per driver is therefore $0.02 × $8,000 = $160 per year.

With only 50 drivers in the pool, actual claims might easily range from 0 to 5, producing total claims anywhere from $0 to $40,000 against expected total losses of $8,000. The volatility is enormous, and the insurer's financial position is dangerously unpredictable.
With 100,000 drivers, the law of large numbers says actual claims will be very close to the expected 2,000, with total costs near $16,000,000. Minor deviations above or below that figure can be covered by a capital reserve without threatening solvency.

The insurer can therefore price each driver's premium at $160 (the pure expected loss) plus a margin for operating expenses and profit. This is only possible because the pool is large enough for the law of large numbers to stabilize predictions.

Prerequisites for the Law of Large Numbers to Work

The law of large numbers works best when three conditions are met:

A large pool: The statistical effect is weak with dozens of risks but powerful with thousands or millions.
Similar, homogeneous risks: Pooling 18-year-old sports car drivers with retired schoolteachers in a single group, and pricing them identically, would produce misleading averages. Insurers create risk classes to group policyholders with similar expected loss rates, so the actuarial predictions are meaningful.
Independence of losses: If one policyholder's loss triggers losses for everyone else in the pool, the averaging effect collapses. A house fire in Chicago is independent of a house fire in Dallas — one does not cause the other. But a hurricane striking a coastal region causes simultaneous losses for every policyholder in that area. Correlated, catastrophic risks (war, pandemic, widespread floods) break the independence assumption, which is why they are often excluded from standard policies or require special treatment like reinsurance.

FAQ

Q: Does a larger pool always produce lower premiums?
A: Not automatically. The law of large numbers reduces variance around the expected loss, but it does not reduce the expected loss itself. If a pool of 1,000,000 people has a higher average expected loss than a pool of 10,000 people, the larger pool will still carry higher premiums — they'll just be more accurately priced. What grows with pool size is predictability and stability, which allows the insurer to operate with smaller safety margins and potentially pass some savings on to consumers.

Q: Why do insurers exclude floods or earthquakes from standard homeowners policies?
A: These perils tend to cause correlated losses — when a flood hits a region, thousands of homes are damaged simultaneously. This violates the independence assumption underlying the law of large numbers, making aggregate losses unpredictable and potentially catastrophic for a single pool. Insurers either exclude these perils entirely, require separate specialized policies, or purchase reinsurance (insurance for insurers) to manage the exposure.