Keeping score in the non-life world
By now you know the moving parts of a P&C insurer: it collects premium, earns it across the policy period, and pays losses that arrive on their own messy schedule. The natural next question is the one every manager, regulator, and investor asks first — *did the business actually work this year?* In life insurance you answer that over decades; in non-life, where contracts last a year, you answer it annually, and you answer it with a handful of simple fractions. This guide is about those fractions: the loss ratio, the expense ratio, and the combined ratio, the three numbers a P&C report leads with.
The whole idea is to ask of every premium dollar: where did it go? Some flowed straight back out as claims. Some paid the people and machinery that run the company. Whatever is left over is the insurer's reward for taking the risk. Each ratio measures one of those leaks as a share of premium, which is why they are expressed as percentages and why they live so naturally side by side. Because every line behaves differently, these ratios are almost always computed by line of business first and only then rolled up — a company can run a beautiful 70% in homeowners while bleeding at 120% in commercial liability.
The loss ratio — and the hidden cost of settling claims
Start with the heartbeat. The [[loss-ratio|loss ratio]] asks: of every premium dollar, how much went straight back out as claims? Collect 100, pay 70 in claims, and the loss ratio is 70%. Precisely, it is incurred losses divided by [[written-vs-earned-premium|earned premium]] for the same period. Two words there carry real weight. *Incurred* means claims belonging to the period whether already paid or still sitting in a reserve — so it captures even the cost of claims not yet settled, not merely the cheques cashed. And the denominator is *earned* premium, matched to the same window of risk; pairing claims with written premium instead would compare apples to a different year's oranges.
But paying a claim is never just handing over money for the damage. Someone must investigate whether the loss is even covered, inspect the crumpled car, hire an engineer to judge a collapsed wall, or pay a lawyer to fight a lawsuit. All those costs of *settling* claims — separate from the loss payment itself — are [[loss-adjustment-expense|loss adjustment expense]], or LAE, and the loss ratio almost always folds them in: loss ratio = (incurred losses + LAE) / earned premium. LAE is not the same as the office overhead and commissions you will meet in the expense ratio; it is part of the cost of the claim. In litigious lines the legal bill can rival the loss itself, so leaving it out would badly flatter the picture.
LAE itself splits into two halves, and the names are worth knowing. Allocated loss adjustment expense (ALAE) is the part that can be traced to a *specific* claim — the outside lawyer or independent appraiser hired for that one lawsuit. Unallocated loss adjustment expense (ULAE) is the overhead that cannot be pinned to any single claim — the salaries of the in-house claims staff, the rent of the claims office, the software that logs every file. If a liability claim pays 100,000 to the injured party plus 30,000 to defence counsel, that 30,000 is ALAE; the slice of the claims department's general running cost attributable to handling it is ULAE. ALAE is usually tracked claim by claim; ULAE is typically estimated as a percentage and spread across the whole book.
The expense ratio and underwriting profit
Claims are only one leak. Running an insurer costs money before a single claim is paid: agents earn commissions, staff draw salaries, governments levy premium taxes, computers and offices cost real cash. The [[expense-ratio|expense ratio]] measures how big a bite all that *running-the-business* cost takes out of premium. If 28 cents of every premium dollar goes to commissions, general operating expenses, and taxes, the expense ratio is 28%. The crucial fence is that these are expenses *other than* the cost of claims — losses and loss adjustment expense live in the loss ratio, never here, and double-counting LAE in both is a classic beginner's error.
One subtlety the loss ratio does not share: the *denominator* of the expense ratio can differ. Acquisition costs like commissions are spent when a policy is *written*, so they are naturally measured against written premium; general administrative costs accrue over time and are often taken against earned premium. Conventions vary by company and country, which means two figures both called 'the expense ratio' can be built on slightly different bases. It is never enough to read the number — you must know which premium base sits underneath it, or you will compare two ratios that are not really speaking the same language.
Subtract both leaks from premium and what remains is the underwriting profit — the money the insurer made purely from the act of insuring, before it earns a penny on its investments. If 100 of premium covers 70 of loss-and-LAE and 25 of expenses, the underwriting profit is 5, an underwriting *margin* of 5%. This is exactly the slice that pricing actuaries plan for in advance: when they build a pure premium up into a gross premium, they deliberately load in an underwriting profit provision so that, if reality matches the assumptions, this very margin appears. The ratios are the scoreboard; the profit provision is the game plan that was supposed to produce that score.
The combined ratio — one number for the whole underwriting result
Now glue the two leaks together. The [[combined-ratio|combined ratio]] is simply the loss ratio plus the expense ratio: it adds the share of premium paid out as claims-and-LAE to the share spent running the company. One number, and a beautifully clear reading of it. Below 100% the underwriting *made* money — premium more than covered everything. Above 100% the underwriting *lost* money — claims and costs ate more than every premium dollar. A 95% combined ratio means 5 of underwriting profit per 100 of premium; a 107% combined ratio means 7 of underwriting *loss* per 100. It is the single most-quoted gauge of how well a P&C insurer did at its core job.
One year on a single line of business (per 100 of earned premium): earned premium 100.0 incurred losses - 68.0 loss adjustment expense (ALAE + ULAE) - 4.0 ---------------------------------------------- loss ratio = (68 + 4) / 100 = 72.0% underwriting expenses (comm, ops, tax) - 25.0 expense ratio = 25 / 100 = 25.0% ---------------------------------------------- combined ratio = 72.0% + 25.0% = 97.0% underwriting profit = 100 - 97 = 3.0 (a 3% margin)
Here is the famous twist, and the single most misunderstood fact in P&C: a combined ratio *above* 100% is not automatically a losing year. The reason is the float. An insurer collects premium today but pays many claims years later — especially in long-tailed liability lines — so in the meantime it holds a great pile of policyholders' money and its own reserves, and it invests them. That investment income is real profit the combined ratio deliberately ignores. If a long-tailed line runs a 105% combined ratio but earns 8% on funds it holds for three years before paying out, the investment return can more than swallow the 5-point underwriting loss, and the line is profitable overall. This is precisely why insurers invest rather than letting reserves sit idle.
Reading the scoreboard honestly
These ratios are simple to compute and easy to misuse, so the craft lies in reading them honestly. First, they are not all settled fact: the loss-ratio numerator leans on reserves, and on a young accident year a chunk of the true cost is still incurred but not reported. A loss ratio that looks heavenly today can deteriorate for years as the losses *develop* toward their ultimate value. A maturing book and a fresh one are simply not comparable at face value.
Second, watch the denominators and the period. Loss ratios use earned premium; expense ratios may mix earned and written; and a fast-growing book — where written premium races ahead of earned — can make the expense ratio look artificially high or low depending on which base is used. Mixing bases without saying so is how two honest analysts arrive at different 'combined ratios' for the same company. Always pin down: which premium, which period, struck at which evaluation date, gross or net of recoveries (salvage and subrogation).
Third, never read a single year in isolation. P&C results swing through the underwriting cycle: years of fierce competition push prices down and combined ratios up, then a wave of losses or a capital squeeze hardens prices and pulls them back. A 103% in a soft market and a 92% in a hard one may say more about where the industry sits in its cycle than about how skilled either insurer is. The ratios are a scoreboard, not a verdict — pair them with the investment return, the line's tail length, and the point in the cycle before you decide whether a year was genuinely good.