Dynamic Pricing and Claims Detection in Motor Insurance
The Actuarial and Commercial Director of an insurer wants to identify which new drivers are highly likely to file serious claims in their first year, in order to adjust the premium in the web quoting tool and attract only profitable profiles.
Where the money comes from
What to do now
- 1Measure the current combined ratio by new-customer segment.
- 2Train the claims model and validate it with last year's portfolio.
- 3A/B test pricing in the web quoting tool in 2 provinces.
For your committee
One page with the decision, the impact, the risks, the steps and the KPI to track.
The print dialog opens: choose "Save as PDF".
Watch out: more volume is not always more profit
Winning more policies by lowering the price for everyone: you attract the risky drivers that the competition rejects (adverse selection). Growing premiums with a worse combined ratio destroys profit.
KPI for the committee
Combined ratio of new business and average premium.
Detail for your data team
Proposed technique: Random Forest
Why: Random Forest combines hundreds of trees that evaluate complex interactions (age + power + area) without overfitting to isolated cases, offering a very solid and stable actuarial scoring.
Cost of getting it wrong: Underestimating risk destroys the insurer's technical margin; overestimating it makes good drivers leave for the competition.
Data needed: Historical Policies: Driver age, years since licence, vehicle power, postcode and use (private/professional).; Claims History: At-fault and not-at-fault claims over the last 5 years (SINCO file).; Telematics / App: Estimated annual mileage and night-driving habits.