BACKGROUND AND CHALLENGE
- Client is a Direct Writer providing P&C insurance to a large member base
- The Homeowners segment of its Property insurance portfolio has performed below industry average
– A key objective was to increase new business, but rates were inadequate
- The challenge was to build tools that could better align price with risk
WHAT WE DID
- Built expected claim loss model that could differentiate policy holders based on claim risk and claim severity.
- Some of the key model variables were:
- Geographical Location
- Previous Claim History
- Statistics Canada (socio-demographic) variables Eg. Education and Occupation
WHAT WAS THE RESULT?
- Each Homeowners Policy was scored and ranked from highest risk (top 10%) to lowest risk (90-100%)
- The Line Chart depicts the percentage of actual losses in the portfolio as predicted by the model (green line) and the current premium (red line) being charged by the Company
- The shaded area represents the “lift”, or increased accuracy in loss prediction provided by the model
- Among the highest risks, the model captures 40% more of the losses than current pricing methods
WHAT WAS THE RESULT STRATEGY?
- Loss Ratio is the amount of claim losses/premium. Listed above is a table that looks at this ratio in terms of establishing a more appropriate pricing strategy for different groups of policyholders