Loss ratio analysis on a stated basis
A loss ratio of ninety-three per cent establishes that a portfolio is under pressure. It does not establish whether a single group has deteriorated, a network has re-tariffed, a benefit has been over-utilised, or the premium has not yet earned. Prizmetrics computes the ratio on a defined basis and attributes the movement to its components.
Last updated 12 September 2026
Definitions
Disagreements about a loss ratio are generally disagreements about its denominator. The definitions are fixed once and carried with every figure derived from them:
- Loss ratio is incurred claims divided by earned net premium.
- Incurred claims comprise paid and outstanding amounts.
- Earned net premium is premium earned pro rata by day to the report date.
The net figure is used because a gross ratio understates the position of a portfolio that cedes its poorest risk. Premium is earned pro rata by day because a policy written in November has not earned a full year of premium by March, and a ratio computed as though it had measures the calendar rather than the portfolio.
Attribution of movement
The ratio is computed for every segment the data supports: group and policy cohort, network and provider, benefit family and claim type, nationality, age band and gender, broker and insurer product.
Each contribution is reported as an absolute value alongside its share of the total. A segment that has moved thirty per cent across two hundred lives and one that has moved three per cent across half the portfolio represent different problems, and a percentage alone does not distinguish between them.
Loss ratio is not additive, and is treated accordingly. The ratio for any row is that row's incurred claims over that row's earned premium; summing the ratios for male and female members produces approximately twice the correct figure. Margins are recomputed rather than summed.
Effect of the report date
Changing the report date does not exclude records. It re-earns premium to the new date, restates exposure, and recomputes every measure derived from either. A comparison between two dates is therefore a comparison of two complete views of the portfolio.
Historical figures consequently change when the report date changes. This reflects the calculation being applied correctly rather than instability in the data, and is the reason every export states the date on which it was computed.
Exclusions from the measure
Claims incurred but not yet reported are not included within incurred claims, as they have not been observed. A loss ratio read without a reserve for them is understated by an amount determined by the proximity of the report date to the period analysed. Estimated IBNR.
Exclusions and data-quality exceptions remain attached to the analysis rather than being resolved silently, so a ratio computed on ninety-six per cent of the claims file is reported as such.