The Law of Large Numbers and Pooling of Risk
One scooter's theft is unpredictable. A lakh scooters' thefts are not. That is why insurance works.
No insurer can tell whether your scooter will be stolen this year. It can tell, quite closely, how many scooters out of a lakh will be. That shift, from an unpredictable single case to a predictable group result, is the law of large numbers, and it is the reason insurance works.
Pooling is the practical side of the same idea: gather many similar risks into one fund, so the losses of the few are paid by the contributions of the many, and the total is stable enough to price.
You save ₹300
- Full 100-question mocks
- Principle-by-principle practice
- Section-wise sets
One payment, no subscription
The Law in Plain Words
The law of large numbers says that as the number of similar, independent exposures grows, the actual result gets closer and closer to the expected result. Toss a coin ten times and you might see seven heads. Toss it ten thousand times and you will be very close to half.
For an insurer, "expected result" means the loss rate found in past data. With a handful of policies, one bad year can be far off that rate. With a very large book, actual claims track the expected rate closely, and the insurer can charge each member a premium that covers the group's losses with confidence.
Why Size Matters: An Illustration
Assume each two-wheeler has a 1 in 100 chance of theft in a year and is worth ₹80,000. The figures are illustrative.
10 bikes
Expected thefts
0.1
What can realistically happen
Usually none, but one theft is a ₹80,000 hit on a tiny fund
Effect on the insurer
Results swing wildly; no stable premium possible
1,000 bikes
Expected thefts
10
What can realistically happen
Somewhere around 10, perhaps a few more or fewer
Effect on the insurer
Predictable enough to price, with a margin
1,00,000 bikes
Expected thefts
1,000
What can realistically happen
Very close to 1,000 in proportion
Effect on the insurer
Losses track expectations closely; the premium can be set near the expected cost
| Pool size | Expected thefts | What can realistically happen | Effect on the insurer |
|---|---|---|---|
| 10 bikes | 0.1 | Usually none, but one theft is a ₹80,000 hit on a tiny fund | Results swing wildly; no stable premium possible |
| 1,000 bikes | 10 | Somewhere around 10, perhaps a few more or fewer | Predictable enough to price, with a margin |
| 1,00,000 bikes | 1,000 | Very close to 1,000 in proportion | Losses track expectations closely; the premium can be set near the expected cost |
Conditions the Law Needs
The prediction only holds if the pool meets these conditions. They reappear as the essentials of an insurable risk.
- check_circleLarge numbers: enough exposures for averages to settle.
- check_circleSimilar (homogeneous) units: houses with houses, young drivers with young drivers. Mixing very different risks blurs the average.
- check_circleIndependence: one loss should not cause the others. A thousand houses on one street can all burn in one fire; a thousand houses spread across a city will not.
- check_circleReliable past data: the expected rate has to come from somewhere, usually years of claims experience or mortality tables.
Quick practice on insurance principles. No signup.
Terms That Go With Pooling
- Pooling
- Combining the risks of many people into one fund, so each pays a share of the group's expected losses rather than bearing their own loss alone.
- Homogeneous exposure units
- Risks similar enough in nature and value to be grouped and priced together.
- Adverse selection
- The tendency of people with higher-than-average risk to buy more insurance. If unchecked, the pool fills with worse risks than the premium assumed. Underwriting and disclosure guard against it.
- Catastrophe (accumulation)
- Many exposures hit by one event, such as a cyclone. It breaks the independence condition, which is why insurers watch concentrations and buy reinsurance.
How IC-01 Tests This
Expect a definition question ("the law of large numbers states that...") and a reasoning one ("insurance is possible mainly because..."). The right answer links large numbers of similar exposures to predictable losses.
The trap is reading the law as "the more policies an insurer sells, the more profit it makes" or "larger numbers reduce the chance of loss for each person." Neither is true. Each person's chance of loss is unchanged; what improves is the insurer's ability to predict the total.
FAQs
What is the law of large numbers in insurance?expand_more
As the number of similar, independent risks in a pool grows, actual losses come closer to the expected losses. That predictability lets an insurer set a fair premium for each member.
What is pooling of risk in insurance?expand_more
Combining many similar risks into one fund. Everyone contributes a premium, and the few who suffer a loss are paid from the fund. The many pay for the losses of the few.
Does the law of large numbers reduce the chance of loss?expand_more
No. Each insured's chance of loss stays the same. What falls is the uncertainty, for the insurer, about the total loss of the group.
Why do insurers need homogeneous risks?expand_more
Because the average only predicts well when the units are alike. Pricing a timber warehouse and a concrete office together would give a rate that suits neither.
Next steps
Take a full IC-01 mock testFull-length IC-01 practice, timed and scored.