AAA - Risk Classification Statement of Principles

I. Summary
Risk classification is the grouping of risks with similar risk characteristics for the purpose of
setting prices.
Risk classification is NOT predicting the experience for an individual risk, identifying
unusually good or bad risks or rewarding groups at the expense of others.
Three primary purposes of the risk classification system:

Protect the insurance system's financial soundness

Be fair

Permit economic incentives to operate and thus encourage widespread availability of

coverage

Five basic principles of the risk classification system:

Should reflect expected cost differences

Should distinguish among risks on the basis of relevant cost-related factors

Should be applied objectively

Should be practical and cost-effective

Should be acceptable to the public

II. Economic Security and Insurance

Mechanisms for coping with the financial impact of chance occurrences:

Hazard Avoidance and Reduction

Practical application is limited, it is wise to transfer major uncertainties

Options for Transfer of Financial Uncertainty

Benefits based on need > Charitable org and Government assistance
Benefits based on contract > Government/Private Insurance, Self-insured group

Public vs Private Programs

Similar > Transfer of financial uncertainty and pooling of risks
Compulsory/Monopoly (Government) vs Voluntary/Competition (Private)

III. The Need for Risk Classification

A. Rational for Risk Classification
Insurance programs can't predict uncertainties but should be able to price them fairly.
Three ways to estimate a price:

Page 1 of 4

AAA - Risk Classification Statement of Principles

Rely on wisdom, insight and good judgement (not the best, only way sometimes)

Observe risk's actual losses over time period (not suitable for life ins., changes in
hazards over time render past information unusable)

Observe groups of similar risks (classes) over time period. Relate price to average
experience of the class. This is the most used approach. The major difficulty is to
select similar risk characteristics before observation period. Optimal set of
characteristics would emerge under perfect competition. In practice, the set reflect
both observed fact and informed judgement.

B. Three Primary Purposes of Risk Classification (FSEFEI)

B1. Protection of Program's Financial Soundness
Adverse selection : Is a threat to solvency. Created by buyers freedom of choice and sellers
inability to select buyers (classify). Risk classification (competition) and restricting buyer's
freedom of choice (monopoly) can control adverse selection.
B2. Enhanced Fairness
Differences in prices among classes should reflect differences in expected costs with no
intended redistribution or subsidy among the classes. Risks in a class should have pretty
much the same expected cost.
B3. Economic Incentive

Offer insurance to as many customers as possible at adequate prices

Intense competition for lower cost risks

Desire to sell to higher cost risks to achieve better market penetration

Increased market penetration provides economies of scale

Competition leads to more refined classification

Additional expense of obtaining more refinement should not be greater than the
reduction in expected costs for the lower cost risk classification (efficiency)

IV. Considerations in Designing a Risk Classification System

A. Underwriting
Underwriting evaluates unique characteristics of each risk. Risk classification is the
framework in which underwriting takes place.
B. Marketing
Influence what products are sold and to whom. Impact the mix of business.
C. Program Design
1. Degree of Choice Available to the Buyer

Compulsory (Broad Classification) VS Voluntary (Refined Classification)

Page 2 of 4

AAA - Risk Classification Statement of Principles

2. Experience Based Pricing

Less refined classification is needed when experience rating adjustments are used

3. Premium Payer

If the buyer of insurance is not the insured, a more broad classification system
may be appropriate since adverse selection potential is reduced (Group ins.)

D. Statistical Considerations
1. Homogeneity

Expected costs for each of the individual risks in a class should be reasonably
similar. It is viewed when the risk is originally classified.

2. Credibility

Classes large enough to allow credible statistical predictions.

3. Predictive Stability

Elements of classification must be useful for predictive purposes, responsive to

changes, yet stable in avoiding abrupt changes in resulting prices.

Increasing the number of classes may improve homogeneity, but at the expenses of
credibility. There is no statistically correct risk classification system.
E. Operational Considerations
1. Expense

Includes obtaining and maintaining the data required to establish classes,

assigning each risk to a class, determining a price for each class

Should be as low as possible for competitiveness and efficiency reasons

Cost of using a variable should be reasonable in relation to benefit achieved

2. Constancy

The lack of constancy in a characteristic's relationship to a risk tends to increase

the expense and reduce the utility of that characteristic, thus reducing the
reliability of the classification system

3. Availability of Coverage

It is desirable for a risk classification system to maximize the availability of ins.

Some risks may be uninsurable. Specific limitation on the coverage available help
to reduce the size of the uninsurable class.

4. Avoidance of Extreme Discontinuities

Establish a reasonable continuum of expected claim costs while having

reasonable differences in prices between classes. Particular care is required for
classes at the extreme ends of the range.
Page 3 of 4

AAA - Risk Classification Statement of Principles

5. Absence of Ambiguity

Definitions of classes should be clear and objective

Classes should be collectively exhaustive and mutually exclusive

6. Manipulation

System should minimize the ability to manipulate or misrepresent a risk's

characteristic

7. Measurability

Characteristic should be susceptible to convenient and reliable measurement

F. Hazard Reduction Incentives

Encourage insured to reduce hazard and expected losses. These incentives are
desirable, but not necessary, features.
G. Public Acceptability
Is hard to apply because social values are difficult to ascertain, vary among segments
of the society and change over time.
Major public acceptability considerations affecting risk classification systems:

They should not differentiate unfairly among risks

They should be based upon clearly relevant data

They should respect personal privacy

They should be structured so that risks tend to identify naturally with their
classification

Laws, regulations and public opinions all constrain risk classification systems.
Legislators must find a balance between public acceptability and market dislocation
due to adverse selection.
H. Causality
Cause and effect relationship increases confidence and public acceptability but it is
often impossible to prove statistically. Causality cannot be made a requirement but
characteristics must be relevant to the insurance provided.
I. Controllability
Refers to ability of a risk to control its own characteristics. Desirable in the context of
hazard reduction incentives and acceptability by the the public, but undesirable in
the context of manipulation where it can render a characteristic irrelevant.

V. Conclusion
Striking the balance is not always easy, but it is in the public interest.

Page 4 of 4