Underlying all of these definitions is the concept of risk pooling, or group sharing, of losses. That is, persons exposed to loss from a particular source combine their risks and agree to share losses on some equitable basis. The risks may be combined under an arrangement whereby the participants mutually insure each other, a plan that is appropriately designated "mutual insurance," or they may be transferred to an organization that, for a consideration called the "premium," is willing to assume the risks and pay the resulting losses. In life insurance, such an organization is a stock life insurance company. While several elements must be present in any sound insurance plan, the essence of the arrangement is the pooling of risks and losses.

The basic principle involved in insurance can best be illustrated in terms of a simple form, such as fire insurance. Suppose that in a certain community there are 1,000 houses, each worth $100,000 and each exposed to approximately the same probability of destruction by fire. The probability that any one of these houses will be destroyed by fire in any particular year is extremely remote, possibly no more than one out of 1,000. Yet if that contingency should occur, the loss to the owner would be staggering�$100,000. If it could be assumed, however, that only one of the 1,000 houses would be destroyed by fire in a particular year, a contribution of only $100 by each home owner would provide a fund large enough to reimburse in full the unfortunate person whose home was lost. If each home owner were willing to assume a certain loss of $100, he could rid himself of the risk of a $100,000 loss. Over the years, only a relatively small percentage of the homes would be destroyed; and through their willingness to contribute a series of small annual sums to a mutual indemnity fund, the property owners would eliminate the possibility of a catastrophic loss to any of their group. The aggregate premium payments over 60 years of home ownership for any one person would still be very small, relative to the protection against the potential loss.

The principle of loss sharing can be applied in identical fashion to the peril of death. The simplest illustration involves insurance for one year, with all members of the group the same age and possessing roughly similar prospects for longevity. The members of this group might mutually agree that a specified sum, such as $100,000, will be paid to the designated beneficiaries of those members who die during the year, the cost of the payments being borne equally by the members of the group. In its simplest form, this arrangement might envision an assessment upon each member in the appropriate amount as each death occurs. In a group of 1,000 persons, each death would produce an assessment of $100 per member. Among a group of 10,000 males aged 35, 21 of them could be expected to die within a year, according to the Commissioners 1980 Standard Ordinary Mortality Table (1980 CSO Table); if expenses of operation are ignored, cumulative assessments of $210 per person would provide the funds for payment of $100,000 to the beneficiary of each of the 21 deceased persons. Larger death payments would produce proportionately larger assessments based on the rate of $2.10 per $1,000 of benefit.

The 1980 CSO mortality table is sex distinct and therefore has different rates at each age for men and women. The rate per $1,000 of benefit for women aged 35 is $1.65, according to the 1980 CSO table. It is very important to note that most large insurance companies base their rates on their own statistics rather than 1980 CSO. The companies that issue policies only to the healthiest applicants will have rates significantly lower than those of the CSO tables used for reserving purposes by the regulators. Even insurance companies issuing policies to applicants in just average health usually experience a rate lower than CSO rates.

Over a century ago, plans based on the assessment technique were widely used in the United States, although confined to fraternal societies and so-called "business assessment associations." Assessments were levied to cover future claims rather than to pay claims that had already been incurred. For example, the Ancient Order of United Workmen, organized in 1868 and the first society to provide death benefits�$2,000 each�levied an assessment of $1 against each member after the payment of each death claim, in order that funds would be available for the prompt settlement of the next claim. Later plans adopted the practice of levying assessments at regular intervals�usually, once a year�rather than after each death.

The early societies generally levied the same assessment on all members, regardless of age. This "flat assessment" plan was based on the theory that there would be a continual flow of new members at the younger ages, with little variation from year to year in the average age of those in the group. Hence the total death rate would not increase, and the annual assessments would remain relatively constant over the years.

Unfortunately, this assumption was invalid. It is not true that the total death rate will not increase so long as the average age of the group does not rise. Suppose, for example, that a fraternal society was organized with 2,000 members, all 40 years of age, and that after several years, its membership was composed of 1,000 persons aged 30 and 1,000 aged 50�an admittedly unrealistic assumption. The average age would still be 40, as it was at the society�s inception. However, since the death rate increases more rapidly from ages 40 to 50 than it decreases from ages 40 to 30, the number of deaths in the group will be greater under the later distribution of ages than under the original. The 1980 CSO Table shows a male death rate of 3.02 per 1,000 at age 40, 1.73 per 1,000 at age 30, and 6.71 per 1,000 at age 50. With 2,000 members aged 40, the society could expect 6.04 deaths in one year, whereas with 1,000 members aged 30 and 1,000 aged 50, it could expect 8.44 deaths. The disparity would have been even larger if a higher average age had been assumed.

Moreover, the average age was virtually certain to increase. Newly organized societies consisted predominantly of young and middle-aged members. Older aged applicants were not solicited, since their admission to the group would have increased the assessments and placed the younger members at a greater financial disadvantage. As the society grew older, however, there was a tendency for the average age to climb because of the difficulty of offsetting the increase in the age of the current membership by the flow of new entrants.

This difficulty can be explained by a simple example. If a society commenced operations with five members aged 20, 21, 22, 23, and 24, the average age of the group would be 22. Assume that during the first year of operation the youngest member, aged 20, dies and is replaced by a new member. If the new member is also 20, the average age of the group will be 22.8, since each of the surviving members is now one year older. If any one except the oldest of the original five members dies and is replaced by a member 20 years of age or more, the average age will increase. The practical effect of this phenomenon is that deceased and withdrawing members of a fraternal society have to be replaced by more than an equivalent number of younger members if the average age of the group is not to increase.

As assessments increased in magnitude and frequency, young and healthy members tended to withdraw from the society, frequently to join a younger society where protection could be obtained at a lower cost, while the old and infirm members remained. This had the obvious effect of increasing the average age even more rapidly, thus further accelerating the withdrawal of the young and healthy members. Under such circumstances it soon became impossible to attract new members. The increase in the proportion of aged and infirm members was accompanied by a corresponding increase in death rates. The inevitable result was an abnormally high rate of assessment and, not infrequently, a collapse of the organization. The attendant loss to those aged members who had all their lives contributed to the benefits of others was disheartening and often tragic.

Once the weakness of the flat assessment plan became apparent, many societies began to grade the assessment according to the age at entry, a typical scale ranging from $0.60 at age 20 to $2.50 at age 60. However, the rate for any given member remained fixed and did not increase as the member grew older and constituted a heavier mortality risk. While not as crude as the flat assessment plan, the graded assessment arrangement proved unsatisfactory and, like the former, worked a hardship upon the younger members.

A third plan called for assessments that would increase as the member grew older. If based on valid mortality data, such increasing premiums were theoretically sound, but from a practical standpoint, the arrangement was defective because it required low premiums in the younger productive years and high premiums in the older years of lessening productive capacity. More serious, it prompted healthy members to withdraw from the plan as premiums increased, lowering the health level of the residual group and producing an abnormal increase in mortality rates. This process is called either antiselection or adverse selection and, while present in many aspects of life insurance and in many different forms, is particularly identified with an insurance plan that has premiums that increase with age.

Finally, some plans provided for a reduction in benefits with advancing age, the assessment rate remaining level. This technique is defensible and is found today in many plans of group life insurance.

As a result of the weaknesses explained above, the assessment plan no longer occupies an important place in the field of life insurance. The plans that had been established on that basis have either become insolvent or have been reorganized in accordance with more commonly accepted principles of life insurance management.