The combined ratio: where an insurer loses premium and how it wins it back

The combined ratio shows how much of its earned premium an insurer hands back — in claims payments and in its own expenses. It is the sum of the loss ratio (claims paid plus the change in reserves, divided by earned premium) and the expense ratio (acquisition and administrative costs over the same base). As long as the ratio stays below the break-even threshold, the insurance business is profitable in its own right; above the threshold, underwriting loses money, and the financial result depends on what the company earns by investing its reserves. For corporate analysis this is the main indicator of the quality of an insurance portfolio: it separates the ability to sell risk from the ability to manage the money that this risk has temporarily left at the company's disposal.

What the ratio is made of and why the base is earned premium

The denominator is not premium collected but premium earned: the part of a written contract that falls within the period already elapsed. If written premium is used instead, rapid sales growth artificially flatters the picture — the money has already come in, while the losses on those contracts have not yet occurred. A simple consequence for analysis follows: for a growing insurer, a combined ratio estimated roughly from premiums collected systematically looks better than the real figure.

The numerator breaks down into two parts that differ in nature. The loss ratio is an estimate, not a fact: it includes reserves for claims that have been reported but not settled and for losses the company does not yet know about. The expense ratio is closer to an observable quantity, but it too involves a choice: whether to recognise commissions paid to intermediaries immediately or as the premium is earned. The logic of allocating costs to a period is the same here as in the accruals ratio, and the attention paid to it is justified: accruals are the channel through which the insurance result is most often smoothed.

Why the ratio can be above the threshold while the company is profitable

An insurer receives the premium before it pays out on losses. The gap between these two moments puts a reserve portfolio at the company's disposal — as a rule a conservative one, built around government debt and high-quality corporate issues. The income from it is not part of the combined ratio and is calculated separately. That is why a business model with a ratio above the threshold exists: the underwriting loss is covered by interest income.

This leads to a rule of reading that matters for corporate finance. The combined ratio cannot be viewed in isolation from interest rates. When money is expensive, weak underwriting is masked by the high yield on reserves; when rates fall, the mask comes off, and the market reprices the company sharply. The portfolio's sensitivity to rates is a separate question, and it should be handled through the discount factor rather than through insurance metrics.

What distorts the indicator

Loss development. The ratio for the current period also includes revisions to estimates for prior years. A release of excess reserves improves it without improving the business; a reserve strengthening does the opposite. Without loss development triangles the two effects cannot be told apart.

Gross or net. The ratio before reinsurance and the ratio after it answer different questions: the first speaks about the quality of underwriting, the second about the risk that has actually remained on the balance sheet. Companies can be compared only on the same basis.

Portfolio structure. Short-tail lines — motor hull, property insurance — give quick feedback: losses are visible almost immediately. Long-tail lines, liability above all, accumulate obligations over years, and the ratio for their early years means little.

Major events. Natural catastrophes, court practice, a one-off change in tariff regulation shift the ratio for a year and then bring it back. The average over the cycle is more useful than the value for the latest period.

How to approach the calculation with Russian data

The source data come from published financial statements — it is convenient to look for them in the reports section and to check disclosure dates in the events calendar. The procedure is as follows. Step 1: take earned premium, not premiums collected. Step 2: add up incurred losses together with the change in reserves. Step 3: single out acquisition and administrative expenses separately, making sure that intermediaries' remuneration is not left outside the calculation. Step 4: repeat the same on a net basis and compare. Step 5: extend the series over several years — an isolated value is useless.

A particular difficulty of the Russian market is that the insurance segment on the exchange is narrow: there are few comparable public insurers with detailed statements available in the list of stocks, and niche players disclose less. The benchmark is therefore more often built not across the market but across the lines of business within the company itself.

What the ratio does not tell you

It says nothing about capital adequacy, about the credit quality of the reserve portfolio or about solvency under stress — for that, the regulator's prudential ratios are needed. Nor does it measure the return to shareholders: consistently strong underwriting usually shows in the ability to pay, which can be checked through the dividend history, but the link here is indirect. Finally, it is not a risk-return metric: an insurer cannot be compared with other investments on this basis; measures such as the Sharpe ratio exist for that purpose, and their definitions are collected in the glossary of terms.

The practical conclusion for corporate valuation: the combined ratio is worth using as a test of underwriting discipline, broken down by line of business and stripped of prior-year loss development, while everything the company earns on its reserves should be moved to a separate line of the analysis. Mixing these two sources of the result is the most common mistake in valuing an insurer.