Interactive workers' comp pricing model
Edit exposures, losses, trend, and expenses to build a live rate indication — the same build-up as the Granite State Mutual case study, fully editable and wired straight into the Decision Studio — or load a real Schedule P workers' comp dataset below.
Interactive WC pricing model
Edit exposures, losses, trend and expenses to build a live rate indication for a workers' comp book, defaulted to the Granite State Mutual case study.
Real data: load a Schedule P dataset
Replace the illustrative Granite State Mutual experience above with a real workers' comp triangle from the CAS Loss Reserve Database (NAIC Schedule P, 1988–1997).
Guided walkthrough
Select a real Schedule P workers' comp dataset and basis, then load it into the model.
| Year | Payroll ($M) | Earned premium ($000) | Rate change % | On-level factor | Exp. weight |
|---|---|---|---|---|---|
| 2017 | 0.980 | ||||
| 2018 | 0.960 | ||||
| 2019 | 0.950 | ||||
| 2020 | 0.970 | ||||
| 2021 | 0.980 | ||||
| 2022 | 0.990 | ||||
| 2023 | 1.000 | ||||
| 2024 | 1.000 |
Class codes
| Class | Description | Payroll ($M) | Current rate/100 | Expected loss rate/100 |
|---|---|---|---|---|
| 8810 | Clerical office employees | |||
| 5645 | Carpentry — dwellings | |||
| 8380 | Automobile service / repair | |||
| 7380 | Drivers & chauffeurs | |||
| 9079 | Restaurant employees (incl. wait staff) |
Assumptions & limitations
- A single blended pure-premium trend (frequency × severity) applies uniformly across all loss dollars over the trend period.
- Large-loss capping is a simplified per-claim proxy (cap × claim count) rather than a full excess-layer simulation.
- Experience-period weights determine both the weighted loss ratio and the credibility claim-count base; adjust them to reflect which years you trust.
- The class rate review caps indicated relativity swings at the selected ± percentage per review cycle, per standard rate-shock management practice.
- This tool is an educational illustration built on the Granite State Mutual synthetic case study, not a real filing.