Price Optimization vs Traditional Ratemaking Techniques (3 differences)
○ Price optimization is a process that uses
§ Big data (data mining of insurance & non-insurance personal info where permitted by law
§ Advanced statistical modeling
○ Price optimization makes granular adjustments to indicated rates
§ Specific risk classifications, or even individual insureds)
○ Traditional ratemaking techniques were judgmental and applied only on a broad level
§ e.g. territory
4 Principles in the CAS “Statement of Principles Regarding Property and Casualty Insurance Ratemaking”
Price Optimization
○ The process of maximizing or minimizing a business metric
○ Uses sophisticated tools and models to quantify business considerations
Cost-based rate
The traditional actuarially derived rate based on loss costs, LAE, and other expenses
Price elasticity of demand
○ The change in quantity demanded versus the price
§ High elasticity –> consumers will shop around even if prices only go up a little (savvy consumers)
§ Low elasticity –> price doesn’t have much effect on demand
Ratebook optimization
Adjust factors in a cost-based rating structure using a demand model
Individual Price Optimization
Build a pricing structure based on both cost and demand
Hybrid Optimization
Insert a new rate factor based on demand (into an existing cost-based structure)
Constrained Optimization
○ Setting minimum and maximum limits on a model’s output
○ Note that unconstrained optimization does not impose these limits
3 Main differences between traditional ratemaking and price optimization
Other differences between traditional ratemaking and price optimization
○ TRADITIONAL RATEMAKING will assign the SAME price to identical risks, but PRICE OPTIMIZATION may assign DIFFERENT prices
○ TRADITIONAL RATEMAKING is generally accepted by regulators whereas PRICE OPTIMIZATION may not be accepted
Benefits of price optimization
○ Provides more accurate pricing
○ If optimization is applied on a ratebook level, it is not unfairly discriminatory
○ Note that individual optimization MAY be unfairly discriminatory
Drawbacks of price optimization
○ Regulators don’t have the data to independently verify rates based on price optimization
○ The models (often GLMs) can produce large individual rate swings (can be controlled by constraints)
○ No evidence of improved stability from using price optimization
○ Concern that ratemaking ASOPs may be violated (if rates are unfairly discriminatory)
Possible regulatory responses to price optimization plans
Disclosures a regulator may require when price optimization is used in a rate filing
○ Rate adjustments that are not cost-based may include judgmental selections)
○ WHETHER price optimization was used
○ WHICH rating factors are affected by price optimization and their quantitative impact
○ WHETHER customers with the same risk profile have different rates
○ Data sources and models that affected the rate charged in any way
Recommendations of the Task Force on Price Optimization regarding pricing methodology
○ Rates should be cost-based
○ Rates should comply with state law
○ Customers with identical risk profiles should be charged the same rate (aside from temporary differences)
Rating considerations that the Task Force believes are unfairly discriminatory
○ Price elasticity of demand
○ Propensity to shop for insurance
○ Retention adjustment at an individual level
○ A policyholder’s propensity to ask questions or file complaints
Recommendations of the Task Force on Pricing Optimization regarding state regulatory practices
○ ISSUE bulletin addressing use of non-cost-based methods
○ ENHANCE disclosure requirements for rate fillings
○ ENSURE compliance with state laws and actuarial principles by analyzing insurer’s rating models