How insurance premiums are priced

A premium is four things stacked on top of each other

When a quote comes back at 1,450 a year, that figure is not one number. It is four, added together and then adjusted for competition.

1. The **pure risk premium**, which is the insurer's honest estimate of what your claims will cost on average.
2. The **expense loading**, which covers the cost of finding you, underwriting you, servicing you and paying your claims.
3. The **capital loading**, which is the return the insurer must earn on the money regulators require it to hold in reserve against the possibility that the estimate is wrong.
4. Minus the **investment offset**, because the insurer holds your premium for months or years before paying claims out of it, and earns something on it meanwhile.

Understanding the stack changes how you read a quote. It explains why two insurers can differ by forty percent on identical risk, why the cheapest quote is sometimes cheap for reasons that will matter at claim time, and why your renewal can rise in a year when nothing about you changed.

It also explains what a premium is not. It is not a prediction that you personally will crash, fall ill or be burgled. It is a price for a slice of a pool.

Layer one, the pure risk premium

Everything starts with one deceptively simple product.

Expected claims cost equals frequency multiplied by severity.

Frequency

Frequency is how often a claim occurs, expressed per policy per year. If 100 policies produce 6 claims in a year, frequency is 0.06, or six percent. Note that this is not "six percent of people crash". It is six claims per hundred policy-years, which is a different statement, because one policyholder can produce two claims and ninety-four can produce none.

Severity

Severity is the average cost of a claim once it occurs. If those 6 claims cost 90,000 in total, average severity is 15,000.

Multiply them. 0.06 multiplied by 15,000 equals 900. That 900 is the pure risk premium: the amount that, collected from every policyholder, would exactly fund the expected claims of the group with nothing left over.

Two consequences follow immediately.

• The pure risk premium is an average, and almost nobody experiences the average. Most people pay 900 and claim nothing. A few claim 15,000 or far more. The pool is the mechanism that turns an unbearable individual outcome into a bearable collective cost.
• Both inputs move independently. Frequency can fall while severity rises, which is exactly what has happened across many motor markets as vehicles become safer but far more expensive to repair, because sensors, cameras and calibration sit in the parts of the car that get hit.

That second point is the single most common reason for premium increases that feel unjustified. Fewer accidents, dearer accidents, and the product of the two can still go up.

Why the pool matters more than you do

Insurers do not price individuals. They price groups and then allocate.

The mathematical reason is the law of large numbers. For a single policy, the outcome is wildly uncertain. Across a hundred thousand similar policies, the average outcome is fairly predictable, and the larger and more homogeneous the group, the tighter that prediction becomes. Predictability is the product being manufactured. The insurer sells you certainty and manufactures it out of scale.

This is why an insurer with a small book in an unfamiliar segment charges more than one with a large book in a familiar segment. It is not greed, it is uncertainty. Uncertainty must be funded, and it is funded through a margin on top of the expected cost.

It is also why your individual claims record matters less than people assume in short-tail personal lines. One claim in three years tells the insurer very little on its own. It matters mostly as a signal, combined with everything else known about the segment you sit in.

Takaful structures approach the same problem from a different direction. Contributions go into a participants' fund from which claims are paid, the operator manages the fund for a fee or a share of surplus rather than owning the underwriting result, and any surplus may be distributed back to participants. The underlying arithmetic of frequency, severity and pooling does not change. What changes is who owns the result and how the operator is paid.

Rating factors are proxies, not verdicts

To allocate the pool's cost fairly, insurers split policyholders into rating cells using factors that correlate with claims experience. Vehicle type, usage, claims history, occupation in some markets, location, sum insured, chosen excess, policy duration and so on.

Three things about rating factors are widely misunderstood.

**They are correlations, not accusations.** A factor earns its place if it separates claims experience across groups, not because an underwriter believes it describes your character. If drivers of a certain vehicle class in a certain area produce more expensive claims on average, that class prices higher, and it prices higher for the careful members of that class too. This is uncomfortable and it is also how pooling works.

**They interact.** The effect of one factor often depends on another, so you cannot reason about a single variable in isolation. Raising your excess might cut ten percent from one quote and two percent from another because the underlying models weight it differently against your other characteristics.

**They are regulated.** Supervisors constrain which factors may be used, how they are disclosed, and how consumers must be treated in the process, alongside broader requirements around solvency and governance Source^source. In the UAE, insurers are licensed and prudentially supervised, with expectations covering actuarial oversight and reserving Source^source.

A cheaper quote is not proof that one insurer thinks you are safer. It is often proof that its model weights one of your characteristics differently, or that it wants growth in your segment this quarter.

The loadings

The pure risk premium of 900 is not a price. It is a cost. Now the loadings go on.

Expenses and acquisition costs

Someone has to answer the phone, run the website, print the schedule, pay the broker commission, investigate the claim and settle it. In personal lines, total expenses commonly consume a meaningful share of premium, and acquisition cost, meaning commission and marketing, is often the largest single component. This is why the same cover through different distribution channels can differ noticeably in price without any difference in the underlying risk.

The cost of capital

Regulators require insurers to hold capital against the risk that claims exceed expectations, on top of technical provisions for claims already expected Source^source. That capital belongs to shareholders or, in mutual structures, to the members, and it must earn a return or it goes elsewhere. So a slice of your premium is effectively rent on the balance sheet standing behind your policy.

The more volatile the line of business, the more capital it consumes and the heavier this loading becomes. Long-tail liability lines carry a far larger capital charge than predictable short-tail lines. That is a large part of why liability cover per unit of protection is priced the way it is.

The risk margin

Accounting frameworks make this explicit. Under IFRS 17, an insurance contract is measured as the expected future cash flows, discounted for the time value of money, plus a separate adjustment for the non-financial risk that those cash flows turn out differently Source^source. That risk adjustment is the formal expression of an old commercial instinct: uncertainty itself has a price, distinct from the expected loss.

Profit margin

Finally, a target underwriting margin. In competitive markets this is frequently thin, and in soft market conditions insurers sometimes accept a negative underwriting result, funding the shortfall from investment income to hold market share.

Worked example, pricing a hypothetical motor book

All figures below are invented for illustration. They are not market rates.

An insurer models a segment of 10,000 similar policies.

1. Expected frequency 0.06, expected severity 15,000, so the pure risk premium is 900 per policy, or 9,000,000 across the book.
2. Expenses run at 25 percent of gross written premium. Note the circularity: expenses are quoted as a percentage of premium, so you cannot simply add them to 900.
3. Target underwriting profit is 5 percent of premium.
4. Investment income is expected to offset roughly 3 percent of premium, because the insurer holds funds between collection and payout.

Set premium P so that P equals 900 plus 0.25P plus 0.05P minus 0.03P. That gives P minus 0.27P equals 900, so 0.73P equals 900, so P is about 1,233.

Now add the capital loading. Suppose the regulator's capital requirement for this book is 4,000,000 and shareholders require a 12 percent return on it. That is 480,000 across 10,000 policies, or 48 per policy, which itself must be grossed up in the same equation. The premium lands near 1,300.

Then the insurer looks at the market, sees competitors quoting 1,450 for comparable cover, and prices at 1,395 to win business while keeping a margin. Or it sees competitors at 1,150, and either accepts a thinner margin or lets the segment go.

Two observations from that example. First, the risk cost was 900 out of roughly 1,300, so about seventy percent of the technical premium. The rest is the cost of running a regulated, capitalised, distributed business. Second, the last step, the market check, is not in any textbook formula, and in competitive lines it moves the final number more than any single rating factor does.

Reserving, and the long delay between price and truth

An insurer sets a price today for claims that will be reported over the next year and paid over the years after that. It will not know whether the price was right for a long time. In the meantime it holds reserves, including a provision for claims incurred but not yet reported.

That delay produces the cycle you can observe in every insurance market. Prices soften while everyone believes reserves are adequate. Then a few years of adverse development emerges, reserves are strengthened, capital tightens, and prices harden across the whole market at once. Individual policyholders experience this as a mysterious industry-wide increase unrelated to their own behaviour, because that is exactly what it is.

Inflation compounds the problem, because it hits severity directly. A claim priced today at 15,000 may settle in two years at 18,000. Insurers must forecast that, and when they forecast it badly in one direction, the correction arrives in later pricing.

Why your renewal rose when you did not claim

Run this diagnostic list before assuming you were singled out.

1. **Claims inflation.** Parts, labour, medical costs and building materials all moved. Severity rose across the pool.
2. **Reserve strengthening.** Prior-year claims developed worse than expected, so the whole book reprices.
3. **Segment deterioration.** Your model of car, your building type or your postcode produced worse results, regardless of your own record.
4. **Loss of an introductory discount.** The first-year price was an acquisition price and was never the technical price.
5. **Sum insured indexation.** The insurer automatically increased your building or contents sum insured, so you are buying more cover than last year.
6. **Reinsurance costs.** Insurers buy insurance too, and when catastrophe reinsurance repriced, that cost flowed into primary premiums.
7. **Capital or interest rate changes.** A change in required returns or discount rates moves the capital loading.
8. **Your own drift.** Age band changed, no-claims discount reached its maximum years ago, mileage estimate revised, a driver added.

None of those are visible on your renewal notice. Ask the insurer or broker which of them applied. A specific answer is a good sign about the firm; a vague one is information too.

Adverse selection, moral hazard, and the defences built into pricing

Two behavioural forces distort any pool, and much of the fine print exists to counter them.

**Adverse selection** is the tendency for those most likely to claim to be most eager to buy. If an insurer priced everyone identically, high-risk applicants would flood in and low-risk ones would leave, driving the average cost up and forcing further increases. Rating factors, medical questionnaires, waiting periods and pre-existing condition clauses are all defences against this spiral. They are not there to catch you out; they are there to stop the pool collapsing.

**Moral hazard** is the tendency to take more risk once insured, or to claim more freely than the loss warrants. Deductibles, co-payments, no-claims discounts, policy limits and claims investigation all push a share of the loss back onto the insured to keep incentives aligned. That is why a policy with no excess at all is priced disproportionately high. You are asking the insurer to absorb both the risk and the behaviour change.

What pricing does not promise

A premium is not a promise that the price is fair to you individually. It is a price that is expected to be adequate across a pool of people who share measurable characteristics with you. If you are the safest driver in your rating cell, you are subsidising others in it. If you are the riskiest, you are being subsidised. Neither fact is knowable in advance by anyone, including you.

A premium is also not a measure of an insurer's willingness to pay claims. Two quotes for the same sum insured can differ in exclusions, sub-limits, claims handling, network access and settlement basis. Compare the schedule of benefits and the exclusions before comparing the price. The cheapest number in a comparison table is a number, not a contract.

And no premium buys certainty about your own outcome. It buys a defined payment, on defined conditions, funded by a pool that must remain solvent long enough to make it.

This article is general financial education. It is not advice about your circumstances and does not recommend any insurer, policy or product.