The four previous lessons dealt with the date of the loss. This one deals with its number, and the question looks scholastic at first: since we know when, what does how many matter. Yet it decides considerable sums, because almost every parameter of a policy is counted per loss. The deductible applies per loss. The limit is expressed per loss and per year. The count is therefore not a quarrel about words: two hundred losses of fifteen thousand euros and one loss of three million produce unrelated settlements from the same facts.
The typical case is a systematic error. A badly configured calculation tool, a template contract containing a defective clause, an erroneous valuation method applied to every file of a financial year: one fault, committed once, and a loss materializing at dozens or hundreds of different clients, on different dates, for different amounts. Each of those clients will write its own letter, and each letter is a claim within the meaning of the contract.
Contracts deal with this by a series clause, whose principle is simple to state: claims arising from one and the same cause are deemed to constitute a single loss. The difficulty is never the principle, it is always the definition of the common cause, and the drafting differs far more than one imagines. Some speak of a single originating event, others of a single error, others of a series of related acts, others again of a common technical cause. Those formulas do not cover the same ground, and on one and the same file they give different counts.
One must then see that a single loss is not systematically the favorable outcome, contrary to assumption. Grouping two hundred claims into one loss means paying one deductible, which is a gain, often considerable where the deductible is high and the claims small. But it opens only one per-loss limit, which is a loss where the total exceeds that limit. The insured therefore has an interest in grouping in one case and in division in the other, and its interest is sometimes the opposite of its insurer's and sometimes the same.
The third effect of the count is attachment in time, and it is the subtlest. If the two hundred claims form a single loss, that loss is deemed to have occurred on a single date, almost always that of the first claim. All the others, including those arriving three years later, attach to that year and therefore to that policy. A professional changing insurer in the meantime will not call on its new policy for the late claims, and that works as much to its advantage, where the old policy was better, as to its detriment.
An underwriting consequence must be added that explains many refusals. An activity producing standardized and repeated work, one template applied to a thousand files, one calculation sold to three hundred clients, carries a risk of a different nature from bespoke work. What concerns the insurer is not the severity of any one file, it is the possibility that a single error replicates automatically. It is also why proposal forms often ask how many clients are served by one and the same method, a question that looks beside the point and is not.
The useful step comes to three readings done together and never separately. Read the series clause and write in one sentence what it treats as the common cause. Read the deductible and ask what, on the company's feared scenario, a count of one loss and then a count of two hundred would produce. And read the limit doing the same exercise in reverse. A professional who has done those three readings knows which side its interest lies on before the discussion opens; one who does them after the first letter discovers its count at the same time as its bill.
An actuarial consultancy sells a standardized pension liability calculation to mid-sized companies. A parameter in its spreadsheet, changed in error in January 2024, produced a systematic understatement of liabilities across 212 files delivered between January and December 2024. The error is discovered by an auditor in March 2026. Losses range from 4,000 to 190,000 euros per client, for an estimated total of 3.4 million euros. The first letters arrive in April 2026, others will keep coming for two years. The policy in force carries a deductible of 25,000 euros per loss and a limit of 2 million per loss and 4 million per year of insurance. The series clause covers "claims resulting from one and the same originating event". One loss or two hundred and twelve?
The clause covers one and the same originating event, and the originating event here is single and perfectly identifiable: a parameter change made once, in January 2024, from which everything else follows mechanically. A count of one loss therefore imposes itself, and what it produces must be measured both ways before rejoicing or complaining. On the deductible, the gain is massive and can be calculated: 212 deductibles of 25,000 euros would come to 5.3 million, more than the total loss of 3.4 million, which would mean the policy served almost no purpose, most files falling below the 25,000 threshold. A single deductible of 25,000 euros is therefore the outcome that saves the file. On the limit, the loss is real but smaller: a single loss opens only the 2 million per-loss limit, whereas the loss reaches 3.4 million, and the consultancy will bear roughly 1.4 million once the deductible is deducted. The 4 million annual limit is of no use here, since it caps an aggregation of losses and there is only one. The net balance therefore favors grouping, which is not the intuition, and it should be verified by calculation rather than assumed. The third effect matters most for what follows and bears on time: the single loss is deemed to have occurred on the date of the first claim, April 2026, and the letters arriving up to 2028 will all attach to the 2026 policy. Two practical consequences follow. The consultancy must not count on its future policies for the late claims, and a change of insurer in 2027 would change nothing in this file. And above all, the 2 million limit being consumed by some of the claims, each new letter arriving afterwards finds a cover already eaten into: the order of arrival decides who is served, exactly as in the extended reporting period. One thing the facts do not state remains to be checked and could move everything: the retroactive date of the policy in force, since the fault is from January 2024 and the policy from 2026.
- 01Almost every parameter of a policy is counted per loss: the deductible, the per-loss limit, so the count decides the amounts.
- 02The series clause groups claims from one common cause, and the whole difficulty lies in the definition of that cause, which varies with the drafting.
- 03A single loss is not systematically favorable: one deductible is a gain, one per-loss limit is a loss.
- 04Grouping also fixes a single date, that of the first claim: late letters attach to that policy and not to later ones.
- 05Standardized repeated work carries a risk of a different nature: it is not the severity of one file, it is the automatic replication of one error.