VITAL STATISTICS AND MORTALITY TABLES. Where, as in fire insurance, policies are issued for a year only or at most for three or five years, it is sufficient to know approximately what the average annual hazard is for the class of risk covered. In fact, fire insurance companies have done with less than this; that is, with merely knowing that, on the whole, the business pays a margin, without knowing that this or that class of hazard has been profitable by itself. If a fire insurance company were, by the nature of the business, compelled to issue policies, without reserving the right to fix the premium on renewal or to refuse to renew, much greater care would, of course, be taken in determining the rates; but even then, the hazard dealt with would not be inherently an increasing one. The problems of classification would per haps be. serious and difficult; but, when they were once solved, the annual risk would not change, the conditions remaining the same.
Life insurance policies usually cover, either for a long term or for the whole period of life, and are issued either for a level premium, or, in any event, for definite premi ums. Even in assessment insurance, where the total amount of assessments per annum may vary, the power is not reserved to vary them as to a particular life, because that life is likely to fail soon. Life insurance cannot be canceled by the company, and it is nearly always renew able at the pleasure of the insured.
If life insurance were issued for a year at a time and the company were free to refuse to renew, the problem would be merely to ascertain in what ratio men, beginning the year in perfect health at each given age, die before its expiration, which could easily be determined by the experience of companies. Of course, in that event the sickly and moribund would be refused the privilege of continuing their insurance.
But the rates of a life insurance company must provide, at age 4o, for instance, not merely for lives newly insured at age 4o, but also for lives which were taken on at age 39 one year ago, at age 38 two years ago, etc. Such lives
will be in an average condition only. Therefore, the statistics of a young company, or of any company with a very large proportion of freshly selected lives, would not be a reliable guide. After the benefit of the selection of the lives by medical examination has worn off, which is estimated to be in about five years, it is usually the case that the insured are at best in no better health than the average of the general population at the same ages; in deed, at the higher ages they may often not average to be in as good condition.
It is evident from the foregoing that great care is requisite to determine just what statistics may be relied upon to gauge the mortality likely to be realized among insured lives. The effects of medical selection, the possi bly persistent effects of selection on account of family history and personal constitutional history, the ill effects of what is known as adverse selection by persons who are conscious of diminished vitality remaining in and others going out more freely, and the disinclination of men to carry insurance to an advanced age, unless convinced that death is near, must all be taken into account. Though, during extended periods, some companies have had very favorable experiences, it is not considered safe or prudent to expect that after the benefits of fresh selection have worn away, the mortality will be materially lower than obtains in the general population.
If, then, an accurate count could be kept of all persons in a nation as they enter upon a certain year of age and of those who die during that year of age, a reasonably reliable death-rate for that age would be the result. This repeated for each age, we would have the death-rate for every age when the work was completed. From this we could readily construct a mortality table, starting with any assumed number of persons, newly born, and dimin ishing that number annually by applying these respective death-rates.