RISK CLASSIFICATION
`
`STATEMENT OF PRINCIPLES
`
`American Academy of Actuaries
`
`Committee on Risk Classification
`
`This booklet has been prepared for an audience generally familiar with insurance concepts and terms
`but not necessarily with the technical aspects of insurance.
`
`
`
`
`
`
`
`
`

`
`TABLE OF CONTENTS
`
`Summary
`
`Economic Security and Insurance
`A.
`Hazard Avoidance and Reduction
`B.
`Transfer of Financial Uncertainty
`C.
`Public and Private Programs
`
`The Need for Risk Classification
`A.
`Rationale for Risk Classification
`B.
`Three Primary Purposes of Risk Classification
`1.
`Protection of Program’s Financial Soundness
`2.
`Enhanced Fairness
`3.
`Economic Incentive
`
`I.
`
`II.
`
`III.
`
`IV.
`
`D.
`
`E.
`
`Considerations in Designing a Risk Classification System
`A.
`Underwriting
`B.
`Marketing
`C.
`Program Design
`1.
`Degree of Choice Available to the Buyer
`2.
`Experience Based Pricing
`3.
`Premium Payer
`Statistical Considerations
`1.
`Homogeneity
`2.
`Credibility
`3.
`Predictive Stability
`Operational Considerations
`1.
`Expense
`2.
`Constancy
`3.
`Availability of Coverage
`4.
`Avoidance of Extreme Discontinuities
`5.
`Absence of Ambiguity
`6.
`Manipulation
`7.
`Measurability
`Hazard Reduction Incentives
`Public Acceptability
`Causality
`Controllability
`
`F.
`G.
`H.
`I.
`
`V.
`
`Conclusion
`
`Page
`1
`
`2
`2
`3
`3
`
`4
`4
`5
`5
`6
`6
`
`7
`7
`8
`8
`8
`8
`9
`9
`9
`10
`10
`11
`11
`11
`12
`12
`13
`13
`13
`13
`14
`14
`15
`
`15
`
`

`
`I.
`
`Summary
`
`Insurance is a means for dealing with the economic uncertainty associated with chance occurrences. It
`does so by exchanging the uncertainty of the occurrence, the timing, and the financial impact of a
`particular event for a predetermined price.
`
`To establish a fair price for insuring an uncertain event, estimates must be made of the probabilities
`associated with the occurrence, timing, and magnitude of such an event. These estimates are normally
`made through the use of past experience, coupled with projections of future trends, for groups with
`similar risk characteristics.
`
`The grouping of risks with similar risk characteristics for the purpose of setting prices is a fundamental
`precept of any workable private, voluntary insurance system. This process, called risk classification, is
`necessary to maintain a financially sound and equitable system. It enables the development of equitable
`insurance prices, which in turn assures the availability of needed coverage to the public. This is
`achieved through the grouping of risks to determine averages and the application of these averages to
`individuals.
`
`It is also important to understand what risk classification is not. Determining average experience for a
`particular class of risk is not the same as predicting the experience for an individual risk in the class. It
`is both impossible and unnecessary to predict experience for individual risks. If the occurrence, timing
`and magnitude of an event were known in advance, there would be no economic uncertainty and
`therefore no reason for insurance.
`
`It is also not the purpose of risk classification to identify unusually good and bad risks or to reward or
`penalize certain groups of risks at the expense of others. Risk classification is intended simply to group
`individual risks having reasonably similar expectations of loss.
`
`Difficulty in risk classification comes with the introduction of concepts such as “fairness” and “similar
`risk characteristics.” Each individual, each business, each piece of property is unique; to the extent that
`the risk classification process attempts to identify and measure every characteristic, it becomes
`unworkable. On the other hand, because there are differences in risk characteristics among individuals
`and among properties which bear significantly upon cost, to ignore all such differences would be unfair.
`Most of the controversy surrounding risk classification involves where the lines are to be drawn.
`
`1
`
`

`
`To achieve and maintain viable insurance systems, the process of risk classification should serve three
`primary purposes. It should:
`!
`!
`!
`
`protect the insurance system’s financial soundness;
`
`be fair; and
`
`permit economic incentives to operate and thus encourage widespread availability of
`coverage.
`
`Striking the appropriate balance among these is not always easy; however, they are clearly in the public
`interest and are not incompatible.
`
`The system should reflect expected cost differences.
`
`The system should distinguish among risks on the basis of relevant cost-related factors.
`
`The following basic principles should be present in any sound risk classification system in order to
`achieve the above purposes:
`!
`!
`!
`!
`!
`
`The system should be applied objectively.
`
`The system should be practical and cost-effective.
`
`The system should be acceptable to the public.
`
`Risk classification is only one factor in an entire set of factors which bear on private, voluntary
`insurance programs. Other factors--such as marketing, underwriting and administration--
`combine with risk classification to provide an entire system of insurance. Changing one factor
`has possible implications on other factors. Changes must be considered in the context of the
`entire system.
`
`II.
`
`Economic Security and Insurance
`
`Society requires various mechanisms for coping with the financial impact of chance
`occurrences, both natural and societal, the prospect of which generates economic insecurity.
`
`A.
`
`Hazard Avoidance and Reduction
`
`Some hazards may be avoided. For example, most of the chance of airplane accidents
`may be avoided by not flying. The incidence and severity of other hazards may be
`
`2
`
`

`
`reduced significantly by taking appropriate safety precautions. For example, the
`installation of smoke detectors or automatic sprinklers may reduce the chance of fire
`losses. However, the practical application of hazard avoidance and hazard reduction is
`limited. Although some financially insignificant hazards may be retained and offset by
`accumulated savings or reserves, the retention of major financial uncertainties may be
`undesirable and unwise. Accordingly, a number of programs which involve a transfer
`of financial uncertainty have been developed.
`
`B.
`
`Transfer of Financial Uncertainty
`
`Programs for transferring financial uncertainty include: sharing among families and
`friends; charitable activities by individuals and organizations; governmental assistance
`and insurance programs; self-insured group pension and welfare plans; and private
`insurance programs.
`
`Certain basic distinctions can be made among these various programs. For example,
`charitable organizations and governmental assistance programs generally provide
`benefits based on demonstrated need, whereas self-insured group pension and welfare
`plans, and governmental and private insurance programs, provide benefits based on
`defined contractual rights.
`
`C.
`
`Public and Private Programs
`
`A comparison of governmental and private programs indicates both similarities and
`differences.
`
`Both types involve the transfer of financial uncertainty from one party to another and the
`subsequent pooling of risks. In both cases, the exposure to loss by the sharing
`mechanism should be broad enough to assure reasonable predictability of the total
`losses.
`
`On the other hand, governmental programs are provided by public law, whereas private
`insurance is provided though an individual contractual arrangement. Governmental
`programs usually are compulsory, while private insurance programs are often voluntary.
`Hence, competition plays a large and vital role in private insurance but little or no role in
`governmental programs. Governmental programs are often devised or needed to
`provide coverage for those hazards which cannot be effectively covered by the private
`insurance system. In governmental programs, the value or cost of benefits received by,
`or paid on behalf of, a class of recipients need not have any long-term relationship to
`the amounts paid into the program by that class. That is contrasted with private
`voluntary insurance programs, where such a long-term relationship is essential.
`
`3
`
`

`
`The private insurance programs are highly diverse. Coverage is available for a wide
`variety of risks, on an individual or group basis, with a variety of underwriting,
`marketing and pricing procedures. As a result, it often difficult to make uniformly
`applicable general statements about private insurance programs.
`
`III.
`
`The Need for Risk Classification
`
`A.
`
`Rationale for Risk Classification
`
`Though an individual exchanges the uncertainty of occurrence, timing and magnitude of
`a particular event for the certainty of a fixed price, that exchange in no way makes the
`uncertain known. Nor need it. The insurance program assuming the financial
`uncertainty is not able to fix the occurrence or, often, the magnitude of a specific risk
`merely because it assumes that risk. But it should find a way of establishing a fair price
`for assuming it.
`
`One way to estimate a price is to rely exclusively on wisdom, insight and good
`judgment concerning the nature of the particular hazard involved and the exposure to
`loss. This usually is not the best method but sometimes is the only one available (as, for
`example, when insuring persons in new occupations which did not exist in the past, or
`persons in unusual occupations for which statistical histories are not meaningful).
`
`A second, theoretically possible way to determine a fair price for the transfer is to
`observe the risk’s actual losses over an extended period of time. This is often not
`appropriate, however. Such an approach offers no solution for risks such as those
`covered by life insurance, where actual observation would show no claims paid while
`the insured individual is alive but an immediate and substantial claim at the time of death.
`Many other risks have this similar characteristic; hindsight suggests there’s little or no
`cost as the individual risk moves to a likely or even certain eventual occurrence. Other
`hazards change so gradually over the period of time needed for the observation that the
`information obtained by observing the past may not be applicable to the current or
`future exposure to financial uncertainty.
`
`A third method is to observe the losses of groups of individual risks with similar risk
`characteristics, which frequently can be done over a shorter period of time. These
`groups are referred to as classes. While any individual risk in a given class is no more
`predictable than it was before the transferring or pooling of the risk occurred, a
`reasonable price may be established by observing the losses of the class and relating the
`price to the average experience of the class. This third approach is the one most often
`used for determining the value of the uncertainty transferred.
`
`4
`
`

`
`A major difficulty with this approach is the need to choose the relevant similar risk
`characteristics and related classes before the observation period. There often is not a
`clear-cut optimal set of characteristics. Over time, in a perfectly competitive market,
`the optimal set of characteristics tends to emerge through the competitive mechanism.
`However, in practice, perfectly competitive markets are seldom achieved, and the risk
`characteristics commonly used reflect both observed fact and informed judgment.
`
`B.
`
`Three Primary Purposes of Risk Classification
`
`A risk classification system serves three primary purposes: to protect the insurance
`program’s financial soundness; to enhance fairness; and to permit economic incentives
`to operate with resulting widespread availability of coverage.
`
`1.
`
`Protection of Program’s Financial Soundness
`
`The financial threat to an insurance program’s solvency is primarily through a complex
`economic concept called adverse selection. It results from the interaction of economic
`forces between buyers and sellers of insurance. In markets where buyers are free to
`select among different sellers, normally with a motivation to minimize the price for the
`coverages provided, adverse selection is possible. In such markets sellers have a
`limited ability to select buyers and have a basic need to maintain prices at a level
`adequate to assure solvency.
`
`In many cases, these economic forces are in equilibrium; occasionally, they are not.
`The freedom of choice and the economic incentive of price may create a dramatic
`movement of buyers to different sellers within an insurance market, or even movements
`into or out of a market. This relocation is the concept of adverse selection, which
`creates economic instability and can threaten the insurance program’s financial stability.
`
`In the early 1900's some assessment societies offered life insurance benefits to
`members without making price distinctions on known mortality differences for different
`age groups. Some younger members of those groups were gradually attracted to lower
`priced competitors, while others decided not to insure at all. This opting out resulted in
`higher prices for remaining members. Some of those remaining then opted out. An
`upward spiral of higher prices resulted for the fewer remaining older lives.
`
`Risk classification is one means of minimizing the potential for adverse selection. It
`reduces adverse selection by balancing the economic forces governing buyer and seller
`actions.
`
`Risk classification is not the only answer to controlling adverse selection. In certain
`
`5
`
`

`
`types of governmental insurance, where participation is mandatory and choices are
`restricted or non-existent (e.g., social security), adverse selection is controlled by a
`restriction of the buyer’s freedom. In a competitive environment, however, risk
`classification is the primary means to control the instability caused by adverse selection.
`
`2.
`
`Enhanced Fairness
`
`Since adverse selection occurs when the prices are not reflective of expected costs, a
`reasonable risk classification system designed to minimize adverse selection tends to
`produce prices that are valid and equitable--i.e., not unfairly discriminatory.
`Differences in prices among classes should reflect differences in expected costs with no
`intended redistribution or subsidy among the classes.
`
`Ideally, prices and expected costs should also match within each class. That is, each
`individual risk placed in a class should have an expected cost which is substantially the
`same as that for any other member of that class. Any individual risk with a substantially
`higher or lower than average expected cost should be placed in a different class.
`
`3.
`
`Economic Incentive
`
`Any economic system that relies primarily on private enterprise for the distribution of
`goods and services relies on companies and individuals to seek out potential customers
`and develop means of successfully selling and servicing the needs of those customers.
`The companies that prove to be the most successful in servicing customers’ needs will
`be rewarded with the largest proportion of the potential customers.
`
`Insurers offering private, voluntary insurance programs are no different in this regard.
`They have incentives to expand their markets and to achieve a high penetration of the
`markets they choose to serve.
`
`In developing marketing strategies, and in pricing the products needed in their markets,
`insurers need a risk classification system that will permit them to offer insurance to as
`many of their potential customers as possible, while at the same time assuring
`themselves that their prices will be adequate to cover the customers’ financial
`uncertainty that they assume.
`
`Generally, competition for the lower cost risks will be the most intense. Therefore,
`prices for these better risks must be different from the prices charged the higher cost
`risks within that market. Also, insurers generally desire to sell insurance to the higher
`cost risks within the same market, in order to achieve better market penetration.
`Increased market penetration provides economies of scale in the marketing or
`
`6
`
`

`
`distribution function, and it also makes it possible for an insurer to provide better
`service to risks in areas where they are more plentiful. Therefore, insurers need the
`ability to price insurance in accordance with the expected costs of each identifiable
`class of risks within their markets.
`
`To be more successful than its competitors would motivate an insurer to become more
`refined in its risk classification system and thus its pricing structure, so that it could serve
`both lower cost and higher cost risks in the marketplace. Thus, there is an incentive for
`risk classifications, as used by competitive insurance programs, to become more refined
`and to more accurately reflect the differentials in expected costs among identifiable
`classes of risks.
`
`Economic incentive also requires the risk classification system to be efficient. The
`additional expense of obtaining more refinement should not be greater than the
`reduction in expected costs for the lower cost risk classification. Thus, there is a
`practical limit to the incentive to add refinements to the classification system.
`
`In general, economic incentive operates over time to favor classification systems that
`result in a price for each risk which most nearly equals the expected cost associated
`with the class to which that risk is assigned.
`
`There is a close, and reinforcing, relationship among these three primary purposes of risk
`classifications. Each is a distinct purpose, yet the system which serves any one tends to serve
`the other two as well.
`
`IV.
`
`Considerations in Designing a Risk Classification System
`
`The ability of any risk classification system to achieve the three described primary purposes is
`substantially influenced by many factors. In particular, this ability is inextricably tied to these
`many design considerations.
`
`A.
`
`Underwriting
`
`Development of an appropriate risk classification system is done without specific regard
`to any of the individual risks to be assumed. It is done a priori and establishes the
`framework within which underwriting can be performed.
`
`Underwriting is the process of determining the acceptability of a risk based on its own
`merits. In contrast to the assignment of a risk to a class based on general criteria, the
`underwriting process involves an evaluation of the individual and possibly unique
`characteristics of each risk.
`
`7
`
`

`
`The design of a risk classification system must recognize that it is applied through the
`underwriting process. In practice, the application of the underwriting function controls
`the practical impact of the classification system, and misapplication of the classification
`system in the underwriting process will achieve results different from those intended.
`
`B.
`
`Marketing
`
`The establishment of a class and a price for that class does not necessarily mean that
`many risks that would be assigned to that class will participate in the insurance program.
`The insurer’s marketing program has an important influence on its mix of business--i.e.,
`what products are sold and to whom. In particular, if those who market private,
`voluntary insurance products are to be held accountable for the program’s ultimate
`economic soundness, arbitrary restrictions on or adjustments to the risk classification
`system by others may produce unintended changes in the mix of business.
`
`C.
`
`Program Design
`
`Certain elements of the design of an insurance program relate quite directly to risk
`classification.
`
`1.
`
`Degree of Choice Available to the Buyer
`
`The design of a risk classification system is affected by the degree to which the
`insurance program is compulsory or voluntary. For programs which are largely
`or entirely compulsory and where there is no voluntary choice among
`competing institutions, broad classifications are sometimes used, the extreme
`being a single class.
`
`Conversely, where participation in the insurance program is voluntary, or where
`there is a voluntary choice among competing institutions or plans, a system that
`classifies risks more broadly than competing systems could invite adverse
`selection.
`
`2.
`
`Experience Based Pricing
`
`Some insurance programs provide for price adjustment after the insurance is
`acquired, based at least in part on the risk’s actual emerging experience.
`
`In the case of insurance purchased by or through an organization, such as an
`employer or association, for a specific group of individuals, this price
`adjustment is referred to as an experience rating adjustment. If the number of
`
`8
`
`

`
`individual risks in the specific group is large enough to produce credible
`experience data, only that groups’ actual experience is used. If the group’s
`data is not adequately credible, its experience is merged with that from other
`comparable groups and the collective experience is used to adjust the price.
`
`In the case of insurance purchased for an individual risk, not grouped risks, the
`price adjustment is made by adjusting the premium paid or by paying a
`dividend. These adjustments are determined by collecting the experience of the
`several individual risks in what is defined as a dividend or equity or experience
`class. The classes used for collecting this experience may or may not be the
`same as the risk classes established and utilized for the original pricing.
`
`To the extent that prices are adjusted based on a risk’s emerging actual
`experience after the insurance and its initial price have been established, less
`refined initial risk classification systems are needed. The experience rating
`refunds, premium adjustments or dividends ultimately produce a refined
`classification system, reflecting at least in part the actual experience of the
`specific risk.
`
`3.
`
`Premium Payer
`
`Under some insurance programs (typically group insurance), the individuals
`insured do not pay the entire price. Such a separation between payer and
`insured can affect the risk classification system in various ways.
`
`If the price is paid by other than the individual insured, the classification system
`is generally a matter of indifference to that individual. It is possible that broad
`classification systems may be appropriate, since the distinction between payer
`and insured can operate to reduce the likelihood of adverse selection.
`
`D.
`
`Statistical Considerations
`
`Risk classification systems are generally based, whenever possible, on statistical
`analysis, modified by informed judgment. Accordingly, certain considerations of a
`statistical nature are involved in designing such a system.
`
`1.
`
`Homogeneity
`
`The expected costs for each of the individual risks in a class should be
`reasonably similar. In a given class, there should be no clearly identifiable
`subclasses with significantly different potential for losses. Significantly dissimilar
`
`9
`
`

`
`risks should be assigned to different classes.
`
`The concept of homogeneity is based upon expected costs as viewed when the
`risk is originally classified. It does not suggest the system can or should
`precisely anticipate the subsequent actual claim experience of a given insured
`risk. The occurrence, timing and magnitude of an unforeseen event for a
`specific risk cannot be predicted in advance. Thus, it is inevitable that not all
`risks in a class will have identical actuarial claim experience. Instead, the
`individual risks’ claim experience will be statistically distributed around the
`average experience for the class. The concept of homogeneity in no way is
`comprised by this inevitable outcome.
`
`By the same token, differences in expected costs between classes do not
`preclude the actual claim experience of some risks in one class from being the
`same as the actual claim experience of risks in another class. This overlap
`phenomenon is both an anticipated and, indeed, statistically inevitable
`ramification of any sound risk classification system.
`
`2.
`
`Credibility
`
`A general statistical principle is that the larger the number of observations, the
`more accurate are the statistical predictions that can be made. Therefore, it is
`desirable that each of the classes in a risk classification system be large enough
`to allow credible statistical predictions about that class. This does not
`necessarily mean that each class must be large enough to stand on its own.
`Accurate predictions for relatively small, narrowly defined classes often can be
`made by appropriate statistical analysis of the experience for broader groupings
`of correlative classes.
`
`3.
`
`Predictive Stability
`
`A major consideration in the construction of risk classification systems and the
`determination of prices for risks in the classes is the prediction of future costs.
`To this end, it is important that elements of a risk classification system be useful
`for predictive purposes. The predictive capability must be responsive to
`changes in the nature of insurance losses, yet stable in avoiding unwarranted
`abrupt changes in resulting prices.
`
`Some statistical tools exist for measuring the historical predictive stability of
`specific risk classification variables. But the actuary must also exercise
`judgment in evaluating noninsurance trends which might reduce the future
`
`10
`
`

`
`effectiveness of predictive power or the practicality of obtaining risk
`classification information. An example of changing predictive value might be
`seen in the recognition of the impact of automobile bumpers meeting certain
`federal safety standards. At one time very few cars had safe bumpers; now,
`most do.
`
`These statistical considerations--homogeneity, credibility and predictive
`stability--are somewhat conflicting. For example, increasing the number of
`classes may improve homogeneity, but at the expense of credibility.
`Consequently, there is no one statistically correct risk classification system. In
`the final analysis, the system adopted will reflect the relative importance
`ascribed to each of these considerations. The decision as to the relative
`weights to be applied will, in turn, be influenced by the nature of the risks, the
`management philosophy of the organization assuming the risk and the judgment
`of the designer of the system.
`
`E.
`
`Operational Considerations
`
`1.
`
`Expense
`
`One important element of a risk classification system is its operational
`expense. These expenses include those for obtaining and maintaining
`the data required to establish classes, for assigning each risk to a class,
`and for determining a price for each class. For reasons of efficiency
`and competitiveness, the expenses should be as low as possible, while
`effectively permitting the system to minimize adverse selection and
`maximize equity.
`
`Further, the cost of utilizing a given variable for classification purposes
`should be reasonable in relation to the benefits achieved, for the
`insurance program and those insured.
`
`2.
`
`Constancy
`
`It is desirable that the characteristics used in any risk classification
`system should be constant in their relationship to a particular risk. This
`constancy should prevail over the period covered by the insurance
`contract or, alternatively, over the period for which a class is assigned.
`This does not preclude the possibility of periodic
`
`11
`
`

`
`reclassification of the risk to take into account changes in the magnitude
`of the classification characteristics. However, the lack of constancy in
`such a characteristic 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 also desirable to provide all of the individuals or groups desiring to
`transfer financial uncertainty the ability to obtain coverage. This means
`that it is desirable to have a classification system which maximizes the
`availability of insurance. To the extent that the classification system
`properly reflects the expected costs of each class and determines the
`price accordingly, overall availability of coverage should be enhanced.
`
`It should be recognized, however, that in some instances the expected
`cost for the highest cost risk class may be of such a magnitude as to
`make the price, from a practical standpoint, unaffordable for some
`insureds. On balance, however, a more refined risk classification
`system properly matching expected cost and price paid will, in the long
`run, enhance rather than inhibit availability of insurance through the
`voluntary market.
`
`There are instances where the risk classification system may actually
`define some risks as necessarily uninsurable. However, even under
`such circumstances it may be possible to minimize the size of the
`uninsurable class by requiring a specific limitation on the coverage
`available to the otherwise uninsurable risk. For example, if an individual
`has been totally disabled by back trouble several times in his life, an
`insurer might require exclusion of disability caused by back trouble from
`coverage as a pre-condition for issuing a new disability policy.
`
`4.
`
`Avoidance of Extreme Discontinuities
`
`There should be enough classes in the system to establish a reasonable
`continuum of expected claim costs but few enough so that differences in
`prices between classes are reasonably significant. Particular attention is
`often required in defining classes at the extreme ends of the range, in
`order to reduce large differences in anticipated average claim costs
`between the extreme class and the adjacent class.
`
`12
`
`

`
`5.
`
`Absence of Ambiguity
`
`The definition of classes should be clear and objective. Once a factual
`assessment of an individual risk has been made, no ambiguity should
`exist concerning the class to which that risk belongs. The classes
`should be collectively exhaustive and mutually exclusive.
`
`6.
`
`Manipulation
`
`The system should minimize the ability to manipulate or misrepresent a
`risk’s characteristics so as to affect the class to which it is assigned.
`
`7.
`
`Measurability
`
`The variables used for classification should be susceptible to convenient
`and reliable measurement. Age, sex, occupation and geographic
`location are examples of factors that are generally reliably determinable.
`Moral character, driving pattern and psychological characteristics are
`examples of factors that are not currently so readily determinable.
`
`F.
`
`Hazard Reduction Incentives
`
`Risk classification systems can be designed to provide incentive for insureds to
`act to reduce expected losses and thus operate to reduce the overall costs of
`insurance in total. For example, recognizing sprinklers for classifying risks for
`fire insurance coverages may encourage their installation and thereby reduce
`expected losses. Or reduced life insurance prices for non-smokers may
`encourage people not to smoke, thus reducing the hazard of premature death
`caused by diseases linked to smoking.
`
`Such incentives are desirable, but not necessary, features of a risk classification
`system. Although worth pursuing, it must be recognized there are limits to
`which a risk classification system can be extended in an attempt to solve
`society’s problems and still serve the necessary and useful purposes for which
`such a system is designed.
`
`G.
`
`Public Acceptability
`
`13
`
`

`
`Any risk classification system must recognize the values of the society in which
`it is to operate. This is a particularly difficult principle to apply in practice,
`because social values
`!
`!
`!
`
`are difficult to ascertain;
`
`vary among segments of the society; and
`
`change over time.
`
`The following are some 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 the risks tend to identify naturally with
`their classification.
`
`Laws, regulations and public opinion all constrain risk classification systems
`within broad social acceptability guidelines. Legislative and regulatory
`restrictions on risk classification systems must balance a desire for increased
`public acceptability with potential economic side effects of adverse selection or
`market dislocation.
`
`H.
`
`Causality
`
`Scientists seek to infer some cause and effect relationship in natural phenomena,
`in order to attempt to understand and to predict. It is philosophically satisfying
`to some when data exhibit such a cause and effect relationship.
`
`Risk classification systems provide a framework of information which can be
`used to understand and predict future insurance costs. If a cause and effect
`relationship can be estab