There are certain fundamental principles that must govern the selection procedures of an insurance company if it is to operate on a sound basis. Some of these principles are mutually inconsistent, which means that a company must fashion its selection in such a manner as to balance these opposing principles.

The range of mortality expectations within which applicants will be regarded as average and hence entitled to insurance at standard rates should be broad enough to encompass the great percentage of applicants. This is particularly important if the company does not offer substandard insurance. An excessive number of rejections undermines the morale of the agency force, increases the cost of doing business, and causes a loss of goodwill among the insurable public. A disproportionate number of substandard policies may have similar effects. Apart from the practical considerations just mentioned, the broader the base of standard risks, the more stable the mortality experience of the group is likely to be. On the other hand, considerations of equity and competition prevent an unwarranted extension of the standard class.

A company must obtain and maintain a proper balance among the risks in each rate classification. This is especially important within the classification of standard risks, which, in view of the principle stated above, is likely to have broad limits. If the overall mortality of the risks in the standard category is to approximate the theoretical average for the group �the goal of most companies�every risk that is worse than average must be offset by one that is better than average. If the range is broad, the margin by which the inferior risk fails to meet the norm for the group should be counterbalanced by the margin by which the offsetting superior risk exceeds the norm. Such precise offsetting or balancing of risks is more of an ideal than an attainable reality. A rough approximation is feasible, however, for a company using the numerical rating system, under which, as explained later, the mortality expectation of individual applications is expressed as a percentage of the average expectation, which is assumed to be 100 percent.

A recent development has constricted the standard range. Companies have developed varying degrees of preferred mortality classes based on such factors as smoking status, tobacco usage in any form, cholesterol level, family history, sports and avocation participation, and other factors. This further refinement of the traditional standard class serves to narrow the remaining standard range not only by removing the superior risks, but also, theoretically at least, by limiting the inferior higher risks allowed in the standard class.

Irrespective of the underwriting procedures used by a company, if each risk classification is overbalanced with risks whose longevity prospects are less favorable than the assumed average for the classification, the company will end up with excessive mortality costs and�unless it enjoys offsetting advantages in other areas of operations�will have difficulty in maintaining its competitive position. The force of this factor will not be diminished by further improving the overall mortality experience of the company unless the rate of improvement is greater than that of its competition �an unlikely situation.

The manner in which applicants are grouped for rating purposes should not unduly violate considerations of equity. Some discrimination among insureds is unavoidable since all risk classifications must be broad enough to include risks of varying quality. Nevertheless the spread between the best and worst risks within a classification should not be so great as to produce rank injustice.

There is a practical side to this matter. If the spread is too great, the better risks may seek insurance with competing companies whose classification system is more equitable, leaving the first company with a disproportionate number of inferior risks. Consequently the first company will have to respond since its premium could be inadequate for the residual group of risks.

The foregoing considerations tend to be relative matters, concerned primarily with equity and competition. There is another factor, however, that operates as an absolute regulator of a company�s underwriting standards�the mortality assumptions entering into the company�s premiums. All mortality tables used by life insurance companies today reflect the experience of insured lives�lives that were subject to some degree of selection. A company�s underwriting standards must be at least as effective as those utilized by the companies that supplied the data for the mortality table. Furthermore the companies that pool their mortality experience for the construction of modern mortality tables employ rather rigorous standards of selection. This is a factor of some importance to companies that, in a desire to capture a larger share of the life insurance market, might be tempted to lower their selection standards. The general improvement in mortality that has been such a prominent feature of the insurance scene during recent decades cannot be expected to nullify the long-run consequences of lax underwriting standards.