HEADNOTE: This is a comprehensive report on critical assumptions and disclosures made in the first required year of implementation of this important standard. The author has compiled data about types of pension plans, funding status, assumptions as to discount rates, future compensation, return on plan assets and other important postulates. He offers suggestions for further improvements in the information required about this significant element of cost. In December 1985, the FASB issued SFAS 87, Employers' Accounting for Pensions, which had been approved by a narrow four to three vote by the Board. This culminated a long process that began in 1974 when FASB added two pension projects to its agenda: 1) accounting and reporting by employee benefit plans; and 2) employers' accounting for pensions. In 1980, SFAS 35, Accounting and Reporting by Defined Benefit Pension Plans, was issued relating to the first project. SFAS 87 is a result of the second project after careful consideration of individual and corporate responses to tentative positions advanced in FASB's Preliminary Views (1982) and Exposure Draft (1985).

As the Board stated, SFAS 87 "continues the evolutionary search for more meaningful and more useful pension accounting." It believes the "Statement represents significant progress, especially in the measurement of net periodic pension cost and in the disclosure of useful information."

Corporations were given the option of adopting most of the provisions of SFAS 87 either in 1986 or 1987. For certain provisions relating to recognition of a minimum liability, plans outside the U.S., and certain smaller employers, an additional year of transition was allowed.

Purpose of Study and Methodology

By the end of 1987, publicly held corporations had implemented the major requirements of SFAS 87. To ascertain the results of selected important disclosures, an analysis was made of 1987 annual reports from a random sample of 130 Standard & Poor's 500 corporations, classified by industry (80 industrial; 20 financial service; 20 utility; 10 transportation). Information extracted from the Financial Statements and other sources was used to answer the following seven questions:

1. What types of pension plans (defined benefit, defined contribution, multi-employer, or profit-sharing) are the companies using?

2. What actuarial assumptions were made relating to discount rates, future compensation rates, and rates of return on assets?

3. How does the selection of the discount rate affect the calculation of the Accumulated Benefit Obligation (ABO)?

4. How do the discount rates reported compare with published Pension Benefit Guaranty Corporation (PBGC) rates for 1987?

5. Are discount rates and rates of return used in the important present value calculations different between under- and over-funded plans?

6. What was the funding status of the plans as measured by Plan Assets/Accumulated Benefit Obligation and Plan Assets/Projected Benefit Obligation ratios?

7. How well is the Plan Asset/Accumulated Benefit Obligation ratio being managed from a liability and future tax planning viewpoint?

Types of Pension Plans

The types of pension plans used by the companies in the sample are summarized in Table 1. A number of companies use more than one type of plan.

As noted, Defined Benefit Plans for single-employers are the most common. Under such plans, the amount of pension benefit to be provided is usually a result of factors such as age, years of service, and compensation.

The second most popular are Defined Contribution Plans. These provide an individual account for each participant and specify how contributions to the individual's account are to be determined. Benefits a participant will receive depend solely on the amount contributed to the account, investment returns on contributions, and forfeitures of other participants' benefits that may be allocated to the account.

As indicated, 13 companies participated in Multi-employer Plans. These plans, in which two or more unrelated employers contribute, usually arise from collective-bargaining agreements. A characteristic of such plans is the "pooling" of assets contributed by the participating employers, with the plan usually administered by a board of trustees composed of management and labor representatives.

The fourth type disclosed are Profit-Sharing Plans. Under such a plan, assets representing a certain percentage of profits are put into a separate trust for retirement benefits or are used to buy annuities or other insurance contracts to provide retirement benefits for eligible employees. All profit-sharing retirement plans reported were for industrial firms.

Actuarial Assumptions

Since the bulk of pension benefits will not be paid for many years, the computation of pension expenses and benefits is based on assumptions about future interest rates, estimates of annual increases in compensation levels, and expected rates of return on plan assets. Prior to SFAS 87, actuarial assumptions as to these rates had to be reasonable overall, but only the expected long-term rate of return on plan assets had to be disclosed. To overcome claims of lack of comparability due to possible widely divergent rates being employed, SFAS 87 requires the disclosure of "the weighted-average assumed discount rate and rate of compensation increase (if applicable) used to measure the projected benefit obligation and the weighted-average expected long-term rate of return on plan assets." Other actuarial assumptions as to mortality, turnover, early retirement, etc., are not separately disclosed. Required disclosure as to rates are summarized in Table 2.

Discount rates are used to reflect the amounts at which pension benefits can be effectively settled. As noted in Table 2, the assumed discount rate for the majority of companies in all industries ranged between 8% and 9.9%, with the average being 8.86%. The range for all plans was from 7% to 10.5%. The lowest average discount rate was 8.47% for the utility companies while the highest was 9.2% for the transportation companies.

The weighted-average expected long-term rate of return on plan assets for all plans was 9.19%. The majority of companies were projecting between 8% and 9.9%, although 27 companies, of 101 reporting, expected above 10%.

The lowest and highest expected rates of return were 8.86% and 9.52%, respectively, for utility and transportation company plans. Since companies in these industries also had the lowest and highest average discount rates, it would appear that utility company management may be slightly more conservative in its evaluation of the future than its transportation industry counterpart.

Annual increases in future salaries and wages for employees covered by defined benefit plans are expected to average 6.03% as shown in Panel C. This projection is significantly above the long-term annual inflation rate of 3% to 3.5%. The highest expected average increase of 6.37% in the financial sector may be due to starting from a lower base and a desire to be more competitive with other industries in attracting and holding qualified personnel. The low average for transportation industry plans of 5.36% is partially attributable to the fact that two of the nine companies reporting expected increases of less than 5%, with one company estimating future annual increases of 1.5%! The author wonders how many of that company's current employees are aware of that projection!

Discount Rate

The importance of the choice of discount rate on accumulated benefit obligations is highlighted by a rule of thumb used by actuaries (Trowbridge, Fair, 1977) that attributes a 20% decrease in such obligations to a 1% increase in the discount rate used in the calculation. The average for all plans was 8.86% and the range was 7% to 10.5% in 1987.

How do these discount rates compare with those published by the Pension Benefit Guaranty Corporation (recommended as one source of information in SFAS 87) in 1987 to establish the value of single- employer plans to be terminated? PBGC rates, calculated five times a year, are very conservative ones that represent, in all likelihood, averages obtained from a sample of life insurance companies to calculate the present value of immediate and deferred annuities. By using conservative rates, present value calculations for employee obligations are higher and, therefore, offer a better retirement benefit valuation base for individual employees.

For example, the range in rates PBGC used in 1987 for immediate annuities (retired employees) was from 7.25% to 8.25% (PBGC, 1988). For deferred annuities (employees still employed), the range was from 4% (annuities commencing at least 15 years hence) to 7.5% (annuities commencing in the next seven years). These rates are substantially below the 8.86% average for all company plans reported in Table 2.

In selecting a discount rate, many pension managers might well use a formula such as the PBGC rates plus 1.5% to 2%. An alternative would be the use of a long-term U.S. Government bond or AAA corporate bond rates, or a combination thereof.

Rates Selected for Under-and Over-Funded Pension Plans

Do actuaries and pension fund managers select higher discount rates and expected rates of return on plans assets, and, hence, lower present values of ABOs (with lower liabilities to be reported) than their counter-parts having over-funded plans? Statistical tests were made separately for discount rates and rates of return on plan assets for both under-and over-funded plans. These tests indicated that there was no significant difference (at the 1% confidence level) in either rate selected for under- or over-funded pension plans.

Funding Status

Companies with defined benefit plans are required to disclose, under SFAS 87, the following: 1) the fair value of plan assets; 2) the accumulated benefit obligation; and 3) the projected benefit obligation. Fair value is "the amount a pension plan could reasonably expect to receive for an investment in a current sale between a willing buyer and a willing seller, that is, other than in a forced or liquidation sale."

The actuarial present value of benefits earned by employees to date is the amount of the Accumulated Benefit Obligation. This amount, less the amount attributable to earned, non-vested benefits, approximates the liability if the plan was to be terminated. The Projected Benefit Obligation is based upon the actuarial present value of benefits earned by employees to date, as well as the future salaries expected to be in effect when benefits are paid to employees. This amount effectively represents the company's obligation as a going-concern.

The Fair Value Plan Asset/Accumulated Benefit Obligation ratio and Fair Value Plan Asset/Projected Benefit Obligation ratio can be thought of as "liquidation" and "going-concern" ratios (Stone, Ingram, 1988) respectively. These ratios are summarized in Table 3.

The data in Table 3 reasonably indicates the financial health for the pension plans as a whole as measured by the liquidation and going- concern ratios discussed above. The financial service firm average was slightly higher for both liquidation and going-concern ratios than the averages for the firms in other industries.

Companies were required to report separately, when they had more than one defined benefit plan, for plans that were "under-funded." "Under- funding" means that the fair value of plan assets was less than the related obligation for that plan. As noted in Table 3, 29 of 131 plans, or 22%, were under-funded on a liquidation basis and 37 of 131 plans, or 28%, were under-funded on a going concern basis. Many of the under- funded plans were relatively small (as measured by assets and liabilities), however, compared with over-funded plans for the same companies.

Liability Management

G.A. David Gray in a recent article (Financial Executive, November/December 1988) strongly suggests that the pension fund manager's relative performance (pension assets outperforming the market) "will give way to an emphasis on outperforming the liability-- specifically, the Accumulated Benefit Obligation (ABO)--and to maintaining and building a pension surplus." The reason for this is that under SFAS 87, pension plan assets must exceed the ABO or the liability will be recorded on the balance sheet.

Another important factor mentioned is that the Omnibus Budget Reconciliation Act of 1988 (OBRA) "prohibits plan sponsors from deducting contributions to pension plans for income tax purposes when the plan is more than 150% funded." Therefore, management of the Plan Asset/ABO ratio is also very important for tax planning.

How well did the companies in our sample do relating to liability management and to the future impact of OBRA? The data in Table 4 provides partial answers by breaking down the aggregative data shown in Table 3 for the important Plan Asset/ABO ratio.

Table 4 indicated that 29 plans, 20 of which were industrial, had ratios of .99 or less which would require reporting a balance sheet liability. However, 102 plans had ratios of 1.0 or higher, indicating a "pension surplus" for those plans with a ratio of greater than 1.0.

From a tax planning viewpoint, 72 of the plans, 46 of which were industrial, had ratios of 1.49 or less and 59 had ratios of 1.5 or higher. Plans with ratios of 1.49 or lower in 1988 and future years, would be normally eligible under OBRA for income tax deductions or contributions to pension plans until the ratio is greater than 1.5.

Summary and Recommendations

Since SFAS 87 requires a standardized method (benefits/years of service) for measuring net periodic pension cost, in contrast to allowing any of five actuarial cost methods acceptable in the past, this required measurement is more understandable, comparable, and hence more useful than those in past practice. Furthermore, expanded disclosures of other important statistical and financial data, as discussed, allow users to understand better the extent and effect of an employer's undertaking to provide employee pensions and related financial arrangements.

The one financial reporting area the author believes can be improved is in narrowing the range of actuarial assumptions relating to rates of discount, rates of return on assets, and rates of compensation increase. Currently each company, and its actuaries, makes its own analysis and determination as to what rates to employ for these variables in the very important present value calculations. As a result, there is perhaps a wider dispersion of rates used than there should be, assuming improved comparability, and hence, usefulness, is a desirable goal.

One example of selection of rates will suffice. Discount rates are to reflect the rates at which pension benefits can be effectively settled. The average discount rate for the 102 companies included in Table 2 of this study, was 8.86% while the range was 7.0% to 10.5%. Eighty of the 102 companies selected a discount rate between 8% and 9.9%. Therefore, it may have been reasonable to require company pension managers to evaluate their own situation but to select a discount rate within the range of, perhaps, 8% to 9.9%. The annual establishing of reasonable ranges for rates of return on assets and rates of compensation increase appears feasible. Such ranges could be established by a committee of the Financial Executives Institute working with a group of actuaries.

An alternative to the establishment of reasonable ranges for key variables would be to require companies to show what the effect of a one-percentage-point change would be on the ABO and on pension cost, holding all other assumptions constant. Such a disclosure is consistent with the FASB Exposure Draft, Postretirement Benefits Other Than Pensions, which calls for disclosure of a one-percentage-point change in the health care trend rate on the measurement of the accumulated postretirement benefit obligation and the postretirement health care benefit cost. Table 1-4 Omitted

Robert H. Mills, PhD, CPA, has been a Professor of Accounting at Lehigh University, Bethlehem, PA, for 25 years. He is a member of the AICPA, PICPA, Accounting Research Association, and a life member of AAA. Dr. Mills has previously contributed articles to The CPA Journal and other professional periodicals.