Contingency income is one of the most over-forecast numbers on an agency's books, and the cause is rarely dishonesty — it's the wrong base. Carriers calculate contingency on earned premium and incurred losses, but agencies often forecast on written premium and paid losses because those numbers are easier to pull. The two methodologies diverge in a growing book and a long-tail book, and they diverge in the seller's favor — which means a buyer who accepts the seller's contingency number is paying for income the carrier's own math won't produce.
§ 01 · Four foundational termsThe numbers that diverge.
| Term | What it measures |
|---|---|
| Written premium | Total premium on policies bound or renewed in the period |
| Earned premium | Premium for coverage actually in force during the period, pro-rated |
| Paid losses | Claim dollars disbursed during the period |
| Incurred losses | Paid plus case reserves plus a provision for unreported claims |
The four terms are easy to conflate and consequential to confuse. Written premium is everything bound or renewed in the period regardless of when coverage starts; earned premium is the pro-rated slice of premium for coverage actually in force during the period. Paid losses are the claim dollars disbursed; incurred losses are paid plus case reserves plus a provision for claims incurred but not yet reported. The two divergences that matter: written exceeds earned in a growing book, and paid is below incurred for long-tail lines. A 10%-growth book has written premium meaningfully above earned — and using written both overstates the qualifying base and understates the calculated loss ratio (the same claims divided by a larger denominator), which doubly flatters the contingency forecast.
§ 02 · Carriers use earned and incurredSo the forecast must too.
Carriers calculate contingency on earned premium and incurred losses, virtually universally — so any forecast built on written premium and paid losses is structurally miscalibrated from the start. Written overstates the qualifying base; paid makes the loss ratio look cleaner than it will settle. The mismatch isn't a rounding error: reconciling a target's modeled contingency to actual carrier statements typically reveals a 10%–25% overstatement.
The miscalibration is systematic, not random, because both shortcuts lean the same direction. Written premium inflates the qualifying base, and paid losses understate the loss ratio against which contingency is tested — so a forecast on written-and-paid is cleaner than the carrier's own calculation will be, every time. When a buyer reconciles a target's modeled contingency against the actual carrier statements and developed reserves, the overstatement typically runs 10%–25%, with a 5%–15% calibration-error band when the methodology shortcuts persist — material in any sub-$3M agency's pro-forma EBITDA. The paid-versus-incurred half of this miscalibration is the subject of paid vs. incurred losses; together the two are the full reconciliation framework.
§ 03 · Three red flagsWhat signals an overstated forecast.
Three M&A red flags reliably surface a contingency forecast built on the wrong base. First, the target's contingency doesn't reconcile to the carrier statements — the number is an estimate, not the actual payment, which is the cleanest tell that a methodology shortcut is in play. Second, a growth-adjusted earned-premium gap — a fast-growing target may use written premium to inflate the forecast, so the faster the growth, the larger the written-versus-earned distortion to check for. Third, reserve development on older periods — current-period loss ratios that haven't been recalculated with developed reserves look artificially clean. The most efficient single pressure-test is the true-up history: clean periodic true-ups signal accurate forecasting, while a pattern of clawbacks signals chronic overstatement that the carrier kept correcting. A buyer who reads the true-up history learns, in minutes, whether the target's contingency number is real or aspirational.
§ 04 · The rebuild and the asksMirroring the carrier's math.
The correct fix is a four-step per-carrier rebuild that mirrors the carrier's own calculation. Pull the contract to identify the exact premium and loss definitions; pull earned premium from the management system or carrier production reports; pull incurred losses (with reserves) from the loss run's incurred column; and reconcile the output against the actual contingency payment for the most recent closed period before projecting forward. The diligence asks that make this possible are three: three years of actual carrier contingency statements (the ones that match the reported P&L revenue), three years of loss runs with both paid and incurred columns plus visible reserve development, and the target's documented contingency-forecasting methodology. The asks are themselves the signal — a target that can't produce all three quickly is running estimates that lean in its favor, and the speed and completeness of the answer reveal the contingency-quality of the book as clearly as the numbers do. Rebuilt on the carrier's definitions, the contingency forecast becomes a number a buyer can actually capitalize. The band the rebuilt number should fall into is in contingency-income benchmarks.
◆
Terminology on this shelf
- Written vs. earned premium
- Bound-in-period versus in-force-during-period — written exceeds earned in a growing book.
- Paid vs. incurred losses
- Disbursed claim dollars versus paid plus reserves plus unreported-claim provision.
- The carrier's basis
- Earned premium and incurred losses — virtually universal, so the forecast must match.
- Overstatement band
- 10%–25% typical overstatement vs. actual statements; 5%–15% calibration error from shortcuts.
- True-up history
- Clean true-ups signal accuracy; a clawback pattern signals chronic overstatement.
- Four-step rebuild
- Pull the contract, earned premium, incurred losses, then reconcile before projecting.