How Insurance Works

In 1688, a group of merchants, ship captains, and underwriters began meeting at Edward Lloyd’s coffeehouse in London to arrange marine insurance. The practice was simple: if your ship sank, several other merchants who had signed the policy would each pay a portion of the loss. No single person bore the catastrophic cost; everyone shared a manageable piece of it. Lloyd’s of London still exists today, it insured David Beckham’s legs for £25 million and the Apollo 11 astronauts couldn’t get life insurance at any price before launch, so NASA had them sign commemorative envelopes that their families could sell if the mission failed. The mechanism underneath all of it, from Lloyd’s coffeehouse to your car insurance app, has barely changed in 300 years.

The short answer

Insurance works by pooling risk across a large group of people. Everyone pays a premium, a regular fee, and the collected premiums pay out claims for the few who suffer a covered loss. The insurer’s job is to predict, with statistical accuracy, how much it will pay out in total, then price premiums so that total income exceeds total claims plus operating costs. What makes this work is the law of large numbers: the more people in the pool, the more predictable the aggregate losses become, even though no individual outcome can be predicted.

The full picture

The law of large numbers: why statistics make insurance possible

A single coin flip is completely unpredictable. Flip a coin 10,000 times and the result will be very close to 5,000 heads. The more flips, the closer you get to exactly 50%.

Insurance applies this principle to loss events. No underwriter can tell you whether your house will burn down this year. But an actuary, a statistician who specializes in risk, can look at 500,000 similar homes and say, with reasonable confidence, that approximately 0.1% of them will experience a significant fire. That’s 500 claims. If the average payout per claim is $40,000, total claims will be approximately $20 million. Divide by 500,000 policyholders and you get the pure mathematical cost of coverage: $40 per policy, before any profit or operating expense.

The precision of these predictions improves dramatically with scale. This is why small insurers are unstable: a bad year with more claims than expected can wipe them out. Large insurers like State Farm, which insures more than 83 million policies according to its 2023 annual report, can absorb unusual claim years that would bankrupt a smaller competitor. The pool size is the safety margin.

Adverse selection: the problem that breaks small pools

Insurance mathematics assumes the risk pool is reasonably representative of the general population. When it isn’t, the math breaks down, a problem economists call adverse selection.

The classic example: if you offer health insurance where sick people can enroll and healthy people can opt out, only sick people will buy it. This drives premiums up, which makes healthy people even less likely to enroll, which makes the pool sicker, which drives premiums higher. Left unchecked, this spiral destroys the insurance market entirely.

Economists George Akerlof and Joseph Stiglitz won the Nobel Prize in Economics in 2001 partly for work formalizing this problem. Akerlof’s famous 1970 paper on “The Market for Lemons” showed how information asymmetry, one party knowing something the other doesn’t, can cause markets to collapse. In insurance, the policyholder usually knows more about their own risk than the insurer does.

Insurers counter adverse selection through underwriting (assessing individual risk before pricing) and mandatory participation (requiring everyone into the pool, the economic logic behind both car insurance requirements and the ACA’s individual mandate).

How premiums are calculated

The basic formula is: Premium = Expected Loss + Loading.

Expected Loss is the statistical prediction: how much the insurer expects to pay out per policy. Loading covers operating expenses, profit margin, and a cushion for prediction error (called a risk load or safety loading).

What makes premium calculation complex is risk classification. Not all policyholders are equally risky. A 17-year-old male driver is statistically far more likely to file an auto insurance claim than a 45-year-old female driver with a clean record. Charging them the same premium would be both financially unstable and unfair to low-risk customers. Insurers use risk factors, age, driving history, claims history, location, credit score, to segment customers into risk categories and price accordingly.

The loss ratio is the metric insurers use to evaluate whether they’ve priced correctly. It’s calculated as: total claims paid ÷ total premiums collected. A loss ratio of 60–70% is generally considered healthy for property and casualty insurance. The remaining 30–40% covers operating expenses and profit. In health insurance, the Affordable Care Act requires insurers to maintain a loss ratio of at least 80% for large group plans, meaning at least 80 cents of every premium dollar must go toward actual medical care.

Reinsurance: insurance for insurers

When a hurricane hits Florida, it can generate tens of billions of dollars in insurance claims simultaneously. No single insurer can absorb that. The solution is reinsurance: insurance companies buy their own insurance from specialized reinsurers.

The global reinsurance market handles roughly $300 billion in premiums annually, according to Swiss Re’s 2024 industry report. The major players, Munich Re, Swiss Re, Hannover Re, exist specifically to absorb catastrophic, correlated losses that would overwhelm primary insurers. When a major hurricane, earthquake, or flood generates unprecedented claims, the losses cascade upward through layers of reinsurance arrangements.

This system is why insurance doesn’t typically fail after natural disasters, at least not immediately. The reinsurance layer provides capital that allows primary insurers to pay claims. The long-term effect shows up in subsequent years, when reinsurance capacity shrinks and primary insurance rates rise in affected regions.

The float: how Berkshire Hathaway built a fortune on insurance

There’s a financial dimension to insurance that most policyholders never consider. You pay your premium in January. Your insurer might pay your claim in October, or never, if you have no losses. In the months between, the insurer holds and invests your money. This pool of held premiums is called the float.

Warren Buffett has described Berkshire Hathaway’s insurance operations as his greatest business insight. As of 2023, Berkshire held over $168 billion in insurance float. If Buffett’s insurance companies operate at break-even on claims, they’re getting hundreds of billions of dollars to invest for free, and the investment returns dwarf the underwriting profits themselves.

Why it matters

The most direct consequence is financial. A higher deductible shifts small losses back to you, small claims are expensive for insurers to process regardless of size, so bearing them yourself is often rational. A $1,000 deductible instead of $500 might reduce your auto premium by 10–15% annually; over five years, that saving often exceeds the extra coverage you gave up.

Bundling auto and home policies saves money for a related reason: the insurer gets more data on your overall risk profile, faces less adverse selection, and has lower acquisition cost. The discount reflects genuine economics.

Globally, insurance penetration, premiums as a percentage of GDP, predicts economic resilience after natural disasters. The 2011 Tōhoku earthquake and tsunami in Japan generated approximately $35 billion in insured losses, a fraction of the $210 billion total economic loss, because earthquake insurance penetration was limited. Economists call this the “protection gap,” and it runs to trillions of dollars globally.

Common misconceptions

“Insurance companies make money by denying claims.” Claim denial is not a reliable profit strategy. Denied claims trigger regulatory review, litigation, and reputational damage that cost more than the claims themselves. Insurers make money through accurate risk pricing, investment of the float, and operating efficiency, not by avoiding legitimate payouts.

“Having insurance means you’ll be made whole after a loss.” Insurance policies are full of exclusions, sub-limits, depreciation calculations, and coverage gaps. A homeowner who assumes their policy covers flood damage will discover, after a flood, that standard homeowner’s insurance does not, the National Flood Insurance Program is a separate federal product. The gap between what people assume their policy covers and what it actually covers is a persistent source of financial devastation after major losses. Reading the declarations page and understanding coverage limits before a claim is the single highest-value thing most policyholders can do.

“Young healthy people are subsidizing older sick people in health insurance.” In community-rated markets (which the ACA created), this is partially true by design. But in employer-sponsored insurance, cross-subsidies run in every direction simultaneously: young subsidize old, healthy subsidize sick, and everyone benefits from the group’s purchasing power. That’s exactly what makes pooling cheaper than individual coverage.