Probable maximum loss (PML) is a modeled estimate of how large catastrophe losses could be at a specified probability level or return period. Insurers and reinsurers use PML to evaluate concentration, capital needs, and the amount of catastrophe protection they purchase.
PML is not a prediction of the next catastrophe. It is a scenario-based risk estimate produced by a catastrophe model under stated assumptions.
The basic idea
An insurer might report a 1-in-100-year or 1-in-250-year PML for a hurricane, earthquake, or other modeled peril. A 1-in-100-year result is often associated with roughly a 1% annual exceedance probability under the model.
That does not mean one such event should occur exactly every 100 years. Return periods describe modeled probabilities, not schedules.
Gross PML versus net PML
The most important comparison is often the basis of the estimate.
- Gross PML measures modeled losses before reinsurance recoveries.
- Net PML measures modeled losses after the effect of qualifying reinsurance and other stated adjustments.
A company can therefore have a very large gross catastrophe exposure while holding a much smaller net retained exposure if its reinsurance program responds as modeled.
Investors should not compare PML figures across insurers unless the peril, geography, return period, model version, gross/net basis, and treatment of reinsurance are reasonably aligned.
PML is model-dependent
Catastrophe models combine assumptions about hazard, property characteristics, vulnerability, insurance terms, event frequency, and loss severity. The resulting PML can change even when the underlying insured portfolio changes only modestly.
Model revisions, updated exposure data, changes in policy terms, geographic mix, inflation, and reinsurance can all move reported PML.
A lower reported PML therefore does not automatically mean the insurer's underlying catastrophe risk became safer.
PML and reinsurance
PML is frequently used to size catastrophe reinsurance. An insurer may purchase protection to cover losses above its Catastrophe Retention up to limits that exceed a selected PML threshold.
This relationship is not exact. Coverage can still differ from modeled loss because of exclusions, sublimits, co-participation, hours clauses, reinstatement terms, aggregate limits, and the fact that actual events need not resemble modeled scenarios.
PML versus Average Annual Loss
PML and Average Annual Loss answer different questions.
- PML focuses on a severe tail scenario at a stated probability level.
- Average annual loss summarizes modeled catastrophe loss across the full distribution over time.
A portfolio can have a high PML but modest average annual loss if extreme events are rare. Another portfolio may have a lower tail PML but more frequent smaller catastrophe losses.
What investors should check
When an insurer discloses PML, ask:
- Is the figure gross or net of reinsurance?
- Which peril and geographic zone does it cover?
- What return period or exceedance probability is used?
- Is it a single-event or aggregate annual measure?
- Which model version and assumptions produced the estimate?
- Does the reinsurance program actually extend above the selected PML?
- How large is the net PML relative to equity or statutory capital?
The ratio of modeled net catastrophe loss to capital can be more informative than the dollar PML alone.
Real-world filing context
Palomar Holdings' 2025 Form 10-K discusses net PML, average annual loss, spread of risk, and catastrophe reinsurance limits when describing its catastrophe-risk management. Everest Group also discusses modeled PML and compares projected net economic loss from catastrophe events with shareholders' equity.
Sources:
Bottom line
Probable maximum loss is a modeled tail-risk estimate, not a forecast. Its usefulness depends on the stated return period, peril, geography, gross-versus-net basis, reinsurance assumptions, and model methodology. Investors should treat PML as one input into catastrophe-risk analysis rather than a complete measure of insurer safety.
Continue Research
Continue from the concept into the Grizzly Bulls research surface that best matches the next question. These links are research continuations, not recommendations or required steps.
Compare catastrophe tail risk
Compare insurer fundamentals alongside stated PML return periods, perils, geographies, gross-versus-net basis, and reinsurance assumptions.
Explore more topics in the Financial Research Encyclopedia.