Estimated IBNR from a stated reporting lag
Claims incurred but not yet reported represent the difference between what a portfolio has cost and what it currently appears to have cost. Prizmetrics estimates that difference from a lag set by the user, applies it only where the conditions permit, and reports the assumption alongside every figure it affects.
Last updated 13 September 2026
Method
Reported incurred claims multiplied by the lag in days divided by 365, applied to the claims within scope, and only where cover ended less than the lag before the report date.
The second condition constitutes half of the rule and is frequently omitted. A policy whose cover ended eight months previously has had eight months in which its claims could be reported; reserving against it as though it remained exposed overstates the portfolio. The run-off condition is the reason the estimate reduces as a cohort matures rather than being carried indefinitely.
The method is a lag model rather than a chain ladder. The simplicity is deliberate: an estimate an underwriter can reproduce independently is one that can be challenged, and a reserve that cannot be challenged is unlikely to be relied upon.
Scope and basis
Reporting behaviour is not uniform across benefits. Inpatient claims are reported late through hospital billing cycles, whereas pharmacy claims are frequently reported same-day. The lag is therefore set per scope rather than once for the portfolio: by benefit family, on the inpatient and outpatient split, or as a single figure where that is appropriate.
The basis is recorded with the assumption, because the outpatient designation refers to two different populations depending on it. The outpatient family of benefit excludes dental, optical and pharmacy; the outpatient setting includes them. A reserve computed on one basis and interpreted as the other is incorrect by the cost of those three benefits.
Disclosure of the assumption
Each Excel workbook and printable report states every lag, its scope, its basis and the run-off condition in full, alongside the report date. An analyst opening the file subsequently can establish what was assumed without further enquiry.
Where no lag is set, the export records estimated IBNR as not calculated rather than leaving a blank field that may be read as zero. A reserve of nil and the absence of a reserve are different statements, and the file distinguishes between them.
Effect on other measures
The reserve is included within incurred claims and therefore affects the loss ratio and the burn cost. Changing a lag restates the portfolio accordingly. The sensitivity of a renewal decision to the reserve assumption is itself material and can be examined before the decision is taken.