Finding the True Cost of Pension Plans

FINANCE

Employee Benefits Plans

Finding the True Cost of Pension Plans

By Thomas T. Amlie

Accounting for defined benefit pension plans has been likened to accounting for incentive stock options with respect to the distorting effects that accounting practice can have on reported income. The recent rise and fall in the equity markets has drawn attention to the way that accounting rules can mask the true costs of the plans from investors. Standard & Poor’s has released a measure of income it calls “core earnings” that would exclude, among other things, any income-increasing measurements of periodic pension cost.

The two chief criticisms of the current accounting methodology involve the resulting “overreporting” of income:

The assets in the pension plan are payable to the beneficiaries of the retirement plan. Because these assets are not the property of the sponsoring business, it is argued that the sponsoring business should not be able to include the return on these assets in the computation of its periodic pension expense (and hence its net income). Instead, the argument is that these increases in asset values are “pledged to” retirees and therefore should not show up as income for the sponsoring business. Some businesses have a return on pension plan assets that dwarfs other components of pension expense. The net result is that pension expense for a period is negative (i.e., pension cost increases reported income).

Businesses often present an overly optimistic picture of their financial condition by using the expected return on pension plan assets, rather than the actual return, when determining their periodic pension cost. Under SFAS 87, Employers’ Accounting for Pensions, this is the appropriate means of computing periodic pension expense. The result, however, is that even though the business’ pension plan assets may have decreased in value over the course of the year, the business still reports a substantial increase in income attributable to its pension plan.

Example: General Electric. General Electric has provided a convenient target for critics of pension accounting. In 2001, GE’s pension plan assets lost $2,876,000,000 in value, due largely to the bear market. GE’s pension expense for fiscal 2001, however, included the expected return on pension plans assets of $4,327,000,000, based on a long-run expected rate of return of 9.5%, which yielded GE a net income from its pension plan of $2,095,000,000.

To many, the GE case embodies all that is wrong with accounting for defined benefit pension plans. GE reports a $2 billion dollar increase in income due to its pension plan, despite the fact that plan assets have lost more than 5% of their value.

The Reasoning Behind Current Standards

The two aforementioned criticisms of pension accounting are easily countered by questioning the nature of accounting for defined benefit pension plans.

Should gains on pension plan assets favorably impact the sponsoring business’s operating results? Over the life of a defined benefit pension plan, the total ultimate cost (or expense) to the sponsoring business will be equal to the total contributions to the plan that the business is required to make. This is the same as any expense that a business incurs: In the long run, the total value of the net resources used is equal to the total expense recorded. In a defined benefit pension plan, each dollar that the pension plan administrator earns on the pension plan assets is one less dollar the business will have to contribute. Therefore, it is only appropriate that returns on pension plan assets serve to decrease net periodic pension cost.

Although some critics might agree in principle with this position, they would argue that recognizing the reduction in pension cost should stop before the business actually shows a gain from its pension operations. Before accepting these arguments, one should consider what conditions must normally exist for a net pension benefit to exist. In order for the expected return on pension plan assets to be greater than the other components of pension expense, the business must have a sizable excess of pension plan assets over pension obligations. (In other possible scenarios, a business might have assets less than or equal to the pension obligation while still reporting a net pension benefit, but these scenarios are highly unlikely.) If businesses are not allowed to recognize the gains on these pension plan assets, they are essentially being penalized for investing their resources in their defined benefit plan rather than in other nonpension investments.

Should businesses use the expected rate of return on pension plan assets in determining their periodic pension cost? In the midst of the current bear market, this criticism has become especially acute. Over the past few years, most businesses have suffered losses on their pension plan assets while continuing to use positive expected rates of return in computing periodic pension costs. Although at first this accounting procedure may seem blatantly misleading, FASB’s reasoning is well founded, even if the ultimate solution is not unassailable.

Defined benefit pension plans involve some of the longest-lived liabilities that a business has. The pension obligation might encompass pension benefits that are expected to be paid out 50 or more years into the future. Because the liabilities associated with the pension obligation have such a long time horizon, it is natural that the assets invested to meet these obligations have a similarly long time horizon. Over such a long time horizon, the appropriate measure of the return on plan assets is the expected long-run average rate of return. In some years (e.g., the bull market of the 1990s), the actual rate of return is substantially greater than the expected long-run average; in other years (e.g., since 2000), the actual rate of return is substantially below the longer-term average. In trying to estimate the long-run total cost of the pension plan (the core problem in accounting for defined benefit pension plans), however, what ultimately matters is the long-run rate of return on plan assets. Determining the periodic pension cost is essentially a process of trying to estimate the total resources the business will need to contribute in order to meet its pension obligation, which will be a function of the long-run rate of return on pension plan assets.

If businesses were required to use the actual rather than the expected rate of return in computing periodic pension cost, short-run market fluctuations would have a substantial and equally misleading impact on reported pension cost. In years where businesses enjoyed substantial above-average returns on plan assets, businesses would report low or negative pension costs, even though in the long run these returns would not persist. Similarly, in years with low returns on plan assets, businesses would be reporting substantially above-average pension costs, even though the increased costs would be a short-term aberration.

Current Treatment of Excess/Deficient Returns on Assets

The current treatment of return on assets under SFAS 87 is essentially an averaging process; periodic pension expense is a function of the long-term expected return on pension plan assets. In some years, the actual return will exceed the expected return; in other years, the actual return will fall short. In the long run, however, the cumulative excess or deficiency in returns should hover around zero. The pattern of actual versus expected return on assets for General Electric for the period of 1991–2001 is presented in the Exhibit.

The annual asset gains or losses are accumulated (along with actuarial gains and losses) and are subject to recognition (amortized to pension cost) only if the cumulative total exceeds a specified threshold. Again, the expectation is that gains in one year will offset losses in other years; the cumulative effect should be nearly zero.

In the current market climate, the low or negative actual returns on pension assets have led critics to argue that businesses are overstating their results by using the expected long-run rate of return. In a bull market environment, where actual returns exceed the long-run average, the corresponding criticism would be that businesses are understating their results by using the comparatively low expected return on assets rather than the higher actual return.

Alternative Procedures

As an exercise in an accounting class, this author asks students to propose alternatives to the current system for accounting for the returns on pension plan assets. While some proposals are off the mark, others are technically viable (although they would seem to unnecessarily confuse financial statement readers). The following are some of the seemingly viable proposed alternatives that have come out of the class:

Use the actual return on plan assets in the determination of pension cost. As noted, this proposal would seem to cause recurring annual misstatements in the true long-run cost of meeting pension obligations. Although it may seem desirable from the perspective of reporting the true operating results of the pension plan for the current period, it would not accurately reflect the expected costs to the business in the long run.

Report the asset gains and losses (i.e., the difference between actual and the expected returns) in the body of the financial statements. Currently, the difference between the expected and actual returns on pension plan assets is not reported in the body of the financial statements; instead, these gains and losses are accumulated off the balance sheet, disclosed in the notes to the financial statements, and subject to recognition only if they exceed a specified amount. The most interesting proposal raised in class has been to treat the cumulative asset gain or loss as an equity item. For example, if there were an asset loss—the expected return on assets was 10% and the actual return was 6%—the expected return would be used in the computation of pension expense, while the 4% asset loss would be booked as a contra-equity charge, with a corresponding 4% unit increase in some measure of the pension liability (e.g., accrued pension cost). Such an approach would leave periodic income unaffected by transitory yearly fluctuations in the asset markets, yet would still disclose, in the body of the financial statements, the actual effects of market return on pension plan assets.

Is Current Disclosure Adequate?

The current disclosures mandated by SFAS 87 provide full and adequate information to those who are capable of understanding them, while the measurement of periodic pension cost represents a reasonable means of attempting to estimate the true long-term cost of providing pension benefits.

In the notes to the financial statements, the actual and expected returns on plan assets are disclosed, as are the various assumptions used in accounting for the defined benefit pension plan. Improving upon these disclosures in some manner may be possible, however. A “reasonably informed” reader of the financial statements probably has little or no education in the fundamentals of accounting for defined benefit plans. Although all of the relevant information related to these plans is disclosed in the notes to the statements, only someone reasonably conversant with the subject would be able to understand the presentation. One way to increase the usability of the information, and preempt criticisms of pension accounting, would be to disclose this information in a presentation more understandable to a layperson.

Thomas T. Amlie, PhD, is an assistant professor of accounting at the State University of New York Institute of Technology, Utica, N.Y.

This Month | About Us | Archives | Advertise| NYSSCPA

The CPA Journal is broadly recognized as an outstanding, technical-refereed publication aimed at public practitioners, management, educators, and other accounting professionals. It is edited by CPAs for CPAs. Our goal is to provide CPAs and other accounting professionals with the information and news to enable them to be successful accountants, managers, and executives in today's practice environments.