Objection to Tables Based on Population Data

For the purposes of measuring the mortality of insured lives, however, it is questionable whether statistics of a general population can be used. Such data, to be sure, would represent the average mortality of a population group and to that extent would approximate the true law of mortality. But for purposes of insurance this may or may not be the mortality rate desired. An insurance company wants a measure of the mortality occurring among insured lives and it is probable that this may differ from that of a specific population group. Insured lives are subject to special influences affecting mortality and these factors must be taken into consideration. The statement has been made that if an insurance company could insure every person who passed a certain corner in a large city until it had a large enough group to guarantee the operation of the law of average, the company could dispense with its medical examination. This is probably true, but the trouble is, when the matter of insurance is left to the choice of the individual, not every one who passed the corner in question would insure; and if this group could be divided into two parts, those who insure and those who do not, the former would show a much higher rate of mortality than the latter. Statistics of insurance companies bear out this statement. Mortality tables based on population statistics formed the first scientific basis for insurance rates, but their approximation to true insurance mortality was not close and they were supplanted by tables based on insured lives as soon as the experience was forthcoming on which to base the latter. The present tables in use by American life-insurance companies and required by most state insurance departments as a basis for the valuation of policy liabilities have been constructed from data of insured lives.