Once all available underwriting information about an insurance application has been assembled, the data must be evaluated and a decision reached as to whether the applicant is to be accepted at standard rates, placed in one of the various substandard classifications, or rejected entirely. This is clearly the focus of the vitally important selection process. Ideally the evaluation and classification system used by a company should (1) accurately measure the effect of each of the factors, favorable and unfavorable, that can be expected to influence an applicant�s longevity; (2) assess the combined impact of multiple factors, including the situations in which the factors are conflicting; (3) produce consistently equitable results; and (4) be simple and relatively inexpensive to operate.

The earliest rating system used in the United States was the "judgment method." Routine cases were processed with a minimum of consideration by clerks trained in the review of applications, and doubtful or borderline cases were resolved by supervisors relying on their experience and general impressions. The method is still used, particularly by smaller companies, and in all insurance companies, the element of judgment plays an important role.

The judgment method of rating functions very effectively when there is only one unfavorable factor to consider or when the decision is simply one of accepting the applicant at standard rates or rejecting the application altogether. It leaves something to be desired when there are multiple unfavorable factors (offset perhaps by some favorable factors) or when the risk, if it does not qualify for standard insurance, must be fitted into the proper substandard classification. To overcome the weaknesses of the judgment method of rating, the numerical rating system, devised over a half century ago, is used today by most insurers.

The numerical rating system is based on the principle that a large number of factors enter into the composition of a risk and that the impact of each of these factors on the longevity of the risk can be determined by a statistical study of lives possessing that factor. It assumes that the average risk accepted by a company has a value of 100 percent and that each of the factors that enter into the risk can be expressed as a percentage of the whole. Favorable factors are assigned negative values, called credits, while unfavorable factors are assigned positive values, called debits. The summation of the debits and credits, added to or deducted from the par value of 100, represents the numerical value of the risk.

Naturally, it would be impossible to assign weights to all the factors that might influence a risk. In practice values are generally assigned to the following 10 factors: (1) build, (2) physical condition, (3) medical history, (4) family history, (5) occupation, (6) aviation and avocation, (7) residence, (8) habits, (9) morals, and (10) plan of insurance.

The values assigned to the various factors are derived from mortality studies among groups of people possessing those characteristics or, in some cases, from estimates of what such mortality studies might be expected to show. For example, if the mortality experience of a group of insured lives with a particular medical history has been found to be 135 percent of that among all standard risks, a debit (addition) of 35 percentage points might be assigned to that medical history. The degrees of extra mortality cited in connection with many of the impairments discussed in the preceding chapter are the basis for the debits under the numerical rating system.

The operation of the system can be illustrated with the following hypothetical case. The applicant is a married man, aged 32, living in Philadelphia, Pennsylvania, with two children. He is 6 feet 1 inch tall and weighs 290 pounds. He is in good physical condition except that, in addition to being overweight, his build is unfavorable�that is, his expanded chest measurement is one inch less than the girth of his abdomen (see chapter 21). His personal health record shows no operations, broken bones, ulcers, or other ailments that would have an adverse effect on longevity. His family is long-lived, and the family history reveals no tuberculosis, insanity, cardiac conditions, malignancies, or diabetes. He has been employed for several years as a warehouseman in an industrial plant. His habits and morals are good. The plan of insurance is 20-payment whole life insurance.

The company might evaluate the facts as follows: The applicant is overweight, which calls for a debit of 50 points according to table 21-1 (see chapter 21). The unfavorable build (girth greater than chest expanded) is the basis for an additional debit of 20 points. The favorable family history receives a credit of 15 points, and the plan of insurance, 20-payment whole life, calls for an additional credit of 10 points. The residence is a neutral factor (only debits are assigned to residence, and debits are usually assessed only for a foreign or tropical residence). Regarding habits and morals, there is no credit for good behavior, only debits for bad behavior. The occupation is also a neutral factor with no debits or credits. Thus the debits add up to 70 points and the credits add up to only 25. Hence the numerical value of the risk is 145.

The analysis is summarized below:

It should be noted that credits are generally not allowed when there are other ratable physical impairments or debits for blood pressure or other cardiovascular-renal impairments.

The ratings obtained by this method may go as low as 75 and as high as 500 or more. The ratings that fall between 75 and 125 are usually classified as standard, although some companies, especially those that do not write substandard insurance, may include risks that produce a rating of 130 (or if the applicant is below age 30, even 140) in the standard category. Risks that produce ratings beyond the standard limit are either assigned to appropriate substandard classifications or declined. The risk is rejected if the company does not write substandard insurance or if the rating is higher than that eligible for the highest substandard classification. Many companies are willing to accept risks that indicate a mortality rate up to 500 percent of normal. Most companies feel that assessing mortality beyond 500 percent will yield results too erratic to price accurately.

As will be explained in the following chapter, the broad category of substandard risks is subdivided into several classifications, each with its own scale of premiums. There may be as few as three or as many as 12 substandard classifications. The full significance of the earlier statements concerning the balancing of risks within each classification should now be apparent. With a spread of 50 percentage points in the standard classification, which comprises the great bulk of accepted risks, it is vitally important that each risk that falls within the 100 to 125 range be balanced (in percentage points) by one that falls in the 75 to 100 category. Otherwise, the average mortality for the entire group will exceed the norm of 100. With a rating of 145, the hypothetical case illustrated above would have fallen into a substandard classification.

The numerical rating system follows much the same procedure as systems that judge a risk without the benefit of numerical values. The same factors are considered, and the final decision is based on the relationship of the various favorable and unfavorable features of the risk. The numerical method, however, sets up objective standards that assist in the final valuation of the risk and allow for greater consistency of treatment. Lay underwriters can process all applications other than those requiring detailed medical analysis and the inevitable borderline cases. This not only expedites the handling of cases but also helps to hold down the expense of the selection process. Underwriters can consult with their home office medical directors on complex medical situations.

Various criticisms, however, have been leveled against the numerical rating system. It has been alleged, for example, that (1) the system is too arbitrary, (2) there are many impairments concerning which knowledge is too limited to permit the assignment of numerical values, (3) the interrelated factors are non- additive in so many cases that it nullifies the value of the numerical process, and (4) too many minor debits and credits are taken into account in evaluating risk. Supporters of the system recognize its flaws but feel that it is still superior to any other method that has been devised. They stress that the procedure must be�and, in practice, is�applied with common sense. They admit that there are many impairments whose effect on longevity cannot be expressed numerically, but they point out that this constitutes a handicap under any method. They argue that under any system, the cases with interrelated impairments require the expert judgment of the insurer�s medical staff and actuarial staff. Whether too many debits and credits are taken into account is a matter of opinion, of course; some companies have modified the numerical rating system to take only major impairments into account.