The methods discussed in this chapter for applying the laws of probability to life insurance have produced a multitude of mortality tables for a variety of purposes. Generally, proprietary company tables are used for premium and dividend calculations. Different tables are used for solvency accounting and reporting to regulators. In the remainder of this chapter we describe important and interesting characteristics of some public tables. Table 13-10 provides a summary of these and other tables developed over the past century.

Two principal kinds of public mortality tables exist. First is the set of United States Life Tables published as a by-product of the decennial census. Census data are used to construct a table for the total population and for each of several sub-populations. The purpose of these tables is to display the public health of the country and to provide a general model for the uncertainty of length of life.

The other principal kind of public table is required by statute in the various states for use by insurance companies for specific purposes. Sections of these codes, which usually follow model laws developed by the National Association of Insurance Commissioners (NAIC), specify how the assets and liabilities on the annual statements of the insurance companies will be calculated. Insurance and annuity contracts in force create liabilities for life insurers. Valuing these liabilities requires reasonable assumptions about several factors, including mortality. The statute specifies a reasonable assumption for the mortality rates of insureds by naming a specific mortality table�the statutory table.

Until 1941 the statutes required tables that were selected from among those in existence in the actuarial profession. The 1941 CSO Table was the first constructed at the request of the NAIC for calculating minimum reserves and surrender values. In keeping with this purpose it was also the first to have an explicit margin added to the mortality rates.

Of the mortality tables listed in table 13-10, those described in rows 3 through 9 are descendants of the 1941 CSO. The 1958 and 1980 CSOs are updated tables used for ordinary life insurance. The 1980 CSO (see table 13-9) differs from its predecessors in several ways. It was the first time actuaries developed and published separate rates for men and women. Another difference was in the treatment of the "hump" in the mortality rates for males around age 20. The raw data in most mortality studies show rates for males reaching a low near age 10, then rising to a maximum near age 21, and falling again to a minimum near age 29. After this, they start a long continuous rise to the end of life. In previous mortality tables, the smoothing process eliminated this hump in the initial rates. In the 1980 CSO, however, because the hump in the male data was more pronounced than ever (due to auto accidents, homicides, drug overdoses, and suicides) it was not eliminated. An examination of table 13-9 shows a hump of 1.91 per 1,000 at age 21, then a reduction to the low of 1.70 at age 28, and then a rise back to 1.91 at age 33. The slight hump apparent in the female data was lost in the smoothing process.

The 1980 CSO also was the first to include selection factors in a table developed for regulatory purposes.

The 1958 and 1980 CET Tables, intended for use with extended term insurance, illustrate the need for mortality tables appropriate to their intended use. There were insufficient data to derive mortality rates for extended term coverage at all ages, so the general level of the mortality was determined and then applied to the pattern of mortality for ordinary life insurance. The resulting tables are based on multiples of the ordinary mortality tables rather than being derived directly from the mortality of lives with extended term coverage.

Further illustrations of tables constructed to be appropriate for the relevant population are the 1961 CSG, CSI, and CIET Tables. These are tables constructed at the request of the NAIC for group, industrial, and industrial extended term policies.

The 1983 Smoker-Nonsmoker tables also were constructed at the request of the NAIC. At the time, companies did not possess sufficient experience on insured lives to construct a table in the usual way. The committee that developed there tables estimated the proportion of smokers at each age of the 1980 CSO population and the relative mortality of smokers versus nonsmokers. Combining this information, the committee estimated mortality rates of the smoking and nonsmoking components of the 1980 CSO populations that were consistent with the overall mortality rates.

The UP-1984 Table was constructed by actuaries to value pension benefits. The Pension Benefit Guarantee Corporation (PBGC) adopted this table to fulfill its mission of protecting defined-benefit pension plan participants with federal guarantees. The PBGC uses this table to value the obligations of pension plans that employers wish to terminate. The table is a unisex table and includes suggestions for complying with laws governing sex discrimination in the distribution of benefits.

Rows 10 through 15 of table 13-10 are tables used for annuity contracts. Only the 1971 and 1983 tables were constructed at the request of the NAIC. The mortality experience of lives covered under group annuities and pensions is different from that for lives covered under individually purchased annuities. These differences indicate a need for separate individual annuitant mortality (IAM) tables and group annuitant mortality (GAM) tables.

All public tables, like those described in this chapter, are intended to provide a minimum basis for valuation. Company solvency is the prime concern. For setting premiums and dividend scales, most major life insurance companies rely on their own recent experience.