-
Adverse Selection
- 1. Company fails to segment business based on meaningful characteristic used by other insurers or does not charge the appropriate diff erential when others do: High-cost insureds select a company due to that company not differentiating these risks from low-cost risks
- 2. Results in distributional shift toward higher-risk insureds for company that doesn't di fferentiate
-
Adverse Selection Process will continue until
- Company improves rate segmentation
- Becomes insolvent
- Decides to focus on high-risk insureds and price accordingly
-
Speed and severity of process depends on various factors
- Whether insureds have full and accurate knowledge of competitor rates
- How much price alone influences purchasing decisions
-
Criteria for evaluating Rating Variables
- 1. Statistical Criteria - rating variables should reflect the variation in expected costs among diff groups: Statistical signi ficance, Homogeneity, Credibility
- 2. Operational Criteria - must be practical to use in rating algorithm: Objective, Inexpensive to administer, Verifiable
- 3. Social Criteria - social acceptability of using a particular risk characteristic: Aff ordability, Causality, Controllability, Privacy
- 4. Legal Criteria - laws and regulations: Statutes, Regulations
-
Distortion of Pure Premium approach to calculate relativities
- Assumes uniform distribution of exposures across all other rating variables
- By ignoring correlation between territory and class, loss experience of various classes can distort the indicated territory relativities: Results in a double-counting e ffect
-
Loss Ratio Approach
- Di fferences from Pure Premium method:
- LR approach uses premium instead of exposure
- LR approach calculates an adjustment to the current relativity
-
Adjusted Pure Premium Approach to calculate relativities
- Adjustment made to Pure Premium approach to minimize impact of any distributional bias
- Use exposures adjusted by the exposure-weighted average relativity of all other variables
- Makes results more consistent with LR method
|
|
