Premium Leakage: The Invisible Twin of Claims Leakage

Claims leakage — overpaying on legitimate claims — gets attention because the money visibly leaves the building. Its twin gets almost none, because the money simply never arrives.

Premium leakage is the gap between the premium you should have charged and the premium you actually charged. It's invisible by construction: there's no payment to audit, no file to review, no anomaly in any report. You just quietly earn less than the risk deserved, policy after policy, for years.

How premium goes missing

Wrong or missing risk characteristics. The rating factor that would have increased the premium was blank, so the system defaulted to an average. Roof age unknown, so the surcharge never applied. Every blank field on a rating-relevant attribute is a silent discount.

Stale values. The property was extended, the fleet grew, the turnover doubled — and your record still reflects the risk as it was at inception. You're insuring today's exposure at yesterday's price.

Misclassification. A commercial risk classified into a cheaper occupancy class because the description was ambiguous and someone picked the nearest match. This one compounds: the misclassification usually persists across every renewal.

Undisclosed exposure. The additional driver, the home business, the extra location. Sometimes deliberate, more often just not asked in a way that captured it.

Discounts that outlive their basis. A no-claims or safety-feature discount applied years ago and never re-verified. Nobody removes a discount; nothing triggers a check.

Rating errors and overrides. An underwriter override to win a deal that quietly became permanent at renewal, carried forward by a system that just copies last year's terms.

Why nobody catches it

Claims leakage at least has a transaction to examine. Premium leakage has an absence — a premium that was never charged, in a policy that looks entirely normal.

To detect it you need a counterfactual: what should this policy have cost, given complete and accurate risk data? And that requires exactly the complete, accurate data whose absence caused the leakage. The problem hides inside its own cause.

So it surfaces indirectly, years later, as a loss ratio that's worse than pricing assumed — and gets attributed to claims inflation, or a bad segment, or luck. Almost never to the fact that a third of the book was rated on incomplete information.

How to actually find it

1. Score rating-field completeness — separately from general data quality. Which fields feed rating, and how complete are they across the in-force book? This single report is usually startling, and most insurers have never produced it. Fields that are 60% complete are 40% guessed.

2. Enrich the in-force book, not just new business. Property databases, aerial imagery, permit records, company registries can fill attributes your systems never captured. Run it retrospectively across in-force policies and compare derived values against recorded ones. The mismatches are your leakage map.

3. Re-rate the book with complete data. Take a sample, enrich fully, re-rate, and compare to charged premium. The aggregate gap — extrapolated — is your premium leakage number. Now it exists as a figure someone can be accountable for, instead of an intuition.

4. Detect change, don't wait for disclosure. External signals — a building permit, a company filing, a registered vehicle change — indicate exposure has shifted. Waiting for the customer to volunteer this is a policy of hoping.

5. Age your discounts and overrides. Every discount should carry an expiry or re-verification trigger. Every underwriter override should be flagged for review at renewal rather than silently inherited.

6. Fix capture at the front, in parallel. Detection recovers the past; better capture prevents the future. Make rating-critical fields required, validated, and — where possible — pre-filled from external data so nobody has to type them.

A word on doing this fairly

This work should be about accuracy, not extraction. Complete data corrects premiums in both directions — plenty of customers are overpaying because a favourable characteristic was never recorded, or a discount they qualify for was never applied.

Framing it as accuracy rather than revenue recovery isn't just ethics; it's practical. It's how you survive the conversation with a regulator, and it's what makes the change defensible to customers and to your own underwriters — who will resist a programme that reads as squeezing the book.

The point

Every insurer scrutinises claims spend, because it's visible. Almost nobody scrutinises whether the premium was right in the first place, because the loss is an absence and absences don't show up in reports.

The uncomfortable arithmetic: a small percentage of premium leakage across an entire book usually dwarfs the claims leakage everyone is chasing — and it's caused by exactly the same thing. Incomplete data at the moment of decision.

Fix the data and both leaks close at once.

We build the enrichment, quality scoring, and unified data that make premium accuracy measurable. More at IntelliBooks.