This is the accessible text file for GAO report number GAO-05-794T entitled 'Private Pensions: Revision of Defined Benefit Pension Plan Funding Rules Is an Essential Component of Comprehensive Pension Reform' which was released on June 7, 2005. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. Testimony: Before the Committee on Finance, United States Senate: United States Government Accountability Office: GAO: For Release on Delivery Expected at 10:00 a.m. EDT: Tuesday, June 7, 2005: Private Pensions: Revision of Defined Benefit Pension Plan Funding Rules Is an Essential Component of Comprehensive Pension Reform: Statement of David M. Walker, Comptroller General of the United States: GAO-05-794T: Mr. Chairman and Members of the Committee: I am pleased to be here today to discuss our recent report on the rules that govern the funding of defined benefit (DB) plans and the implications of those rules for the problems facing the Pension Benefit Guaranty Corporation (PBGC) and the DB pension system generally.[Footnote 1] In recent years, the PBGC has encountered serious financial difficulties. Prominent companies, such as Bethlehem Steel, U.S. Airways, and United Airlines, have terminated their pension plans with severe gaps between the assets these plans held and the pension promises these plan sponsors made to their employees and retirees. These terminations, and other unfavorable market conditions, have created large losses for PBGC's single-employer insurance program- -the federal program that insures certain benefits of the more than 34 million participants in over 29,000 plans. The single-employer program has gone from a $9.7 billion accumulated surplus at the end of fiscal year 2000 to a $23.3 billion accumulated deficit as of September 2004, including a $12.1 billion loss for fiscal year 2004. In addition, financially weak companies sponsored DB plans with a combined $96 billion of underfunding as of September 2004, up from $35 billion as of 2 years earlier. Because PBGC guarantees participant benefits, there is concern that the expected continued termination of large plans by bankrupt sponsors will push the program more quickly into insolvency, generating pressure on the Congress, and ultimately the taxpayers, to provide financial assistance to PBGC and pension participants. Given these concerns, we placed the PBGC's single-employer program on GAO's high-risk list of agencies and programs that need broad-based transformations to address major challenges. In past reports, we identified several categories of reform that the Congress might consider to strengthen the program over the long term. We concluded that the Congress should consider comprehensive reform measures to reduce the risks to the program's long-term financial viability and thus enhance the retirement income security of American workers and retirees.[Footnote 2] More broadly, pension reform represents a real opportunity to address part of our long-term fiscal problems and reconfigure our retirement security systems to bring them into the 21st century.[Footnote 3] This opportunity has many related pieces: addressing our nation's large and growing long-term fiscal gap; deciding on the appropriate role and size of the federal government--and how to finance that government--and bringing the wide array of federal activities into line with today's world. Continuing on our current unsustainable fiscal path will gradually erode, if not suddenly damage, our economy, our standard of living, and ultimately our national security. We therefore must fundamentally reexamine major spending and tax policies and priorities in an effort to recapture our fiscal flexibility and ensure that our government can respond to a range of current and emerging security, social, economic, and environmental changes and challenges. The PBGC's situation is an excellent example of the need for the Congress to reconsider the role of government organizations, programs, and policies in light of changes that have occurred since PBGC's establishment in 1974. PBGC's challenges bear many similarities to the challenges facing our Social Security system. Both programs have adequate current revenues and assets to pay promised benefits for a number of years; yet, both face large and growing accumulated deficits on an accrual basis. As a result, timely action to address both private pension and Social Security reform is needed. In pursuing such reforms, consideration should be given to the interactive effects of any such reforms and how they contribute to addressing our nation's large and growing fiscal challenge, key demographic, economic and workforce trends, and the economic security of Americans in their retirement years. Our recent work on DB pension funding rules provides important insights in understanding the problems facing PBGC and the DB system. To summarize our findings, while pension funding rules are intended to ensure that plans have sufficient assets to pay promised benefit to plan participants, significant vulnerabilities exist. Although from 1995 to 2002 most of the 100 largest DB plans annually had assets that exceeded their current liabilities, by 2002 over half of the 100 largest plans were underfunded, and almost one-fourth of plans were less than 90 percent funded.[Footnote 4] Further, because of leeway in the actuarial methodology and assumptions that sponsors may use to measure plan assets and liabilities, underfunding may actually have been more severe and widespread than reported. Additionally, on average over 60 percent of sponsors of these plans made no annual cash contributions to their plans. One key reason for this is that the funding rules allow a sponsor to satisfy minimum funding requirements without necessarily making a cash contribution each year, even though the plan may be underfunded. [Footnote 5] Further, very few sponsors of underfunded plans were required to pay an additional funding charge (AFC), a funding mechanism designed to reduce severe plan underfunding. Finally, our analysis confirms the notion that plans sponsored by financially weak firms pose a particular risk to PBGC, as these plans were generally more likely to be underfunded, to be subject to an additional funding charge, and to use assumptions to minimize or avoid cash contributions than plans sponsored by stronger firms. Background: In DB plans, formulas set by the employer determine employee benefits. DB plan formulas vary widely, but benefits are frequently based on participant pay and years of service, and typically paid upon retirement as a lifetime annuity, or periodic payments until death. Because DB plans promise to make payments in the future and because tax- qualified DB plans must be funded, employers must use present value calculations to estimate the current value of promised benefits.[Footnote 6] The calculations require making assumptions about factors that affect the amount and timing of benefit payments, such as an employee's retirement age and expected mortality, and about the expected return on plan assets, expressed in the form of an interest rate. The present value of accrued benefits calculated using mandated assumptions is known as a plan's "current liability." Current liability provides an estimate of the amount of assets a plan needs today to pay for accrued benefits. The Employee Retirement Income Security Act of 1974 (ERISA), and several amendments to the law since its passage, established minimum funding requirements for sponsors of pension plans in order to try to ensure that plans have enough assets to pay promised benefits. Compliance with the minimum funding requirements is recorded through the plan's funding standard account (FSA). The FSA tracks events that affect the financial health of a plan during that plan year: credits, which reflect improvements to the plan's assets, such as contributions, amortized experience gains, and interest; and charges, which reflect an increase in the plan's financial requirements, such as the plan's normal cost and amortized charges such as the initial actuarial liability, experience losses, and increases in a plan's benefit formula.[Footnote 7],[Footnote 8] ERISA and the Internal Revenue Code (IRC) prescribe rules regarding the assumptions that sponsors must use to measure plan liabilities and assets. For example, for plan years 2004 and 2005, the IRC specifies that the interest rate used to calculate a plan's current liability must fall within 90 to 100 percent of the weighted average of the rate on an index of long-term investment-grade corporate bonds during the 4- year period ending on the last day before the beginning of the plan year.[Footnote 9] Similarly, rules dictate that sponsors report an "actuarial" value of assets that must be based on reasonable assumptions and must take into account the assets' market value. This value may differ in any given year, within a specified range, from the current market value of plan assets, which plans also report. While different methodologies and assumptions will change a plan's reported assets and liabilities, sponsors eventually must pay the amount of benefits promised; if the assumptions used to compute current liability differ from the plan's actual experience, current liability will differ from the amount of assets actually needed to pay benefits.[Footnote 10] Funding rules generally presume that the plan and the sponsor are ongoing entities, and plans do not necessarily have to maintain an asset level equal to current liabilities every year. However, the funding rules include certain mechanisms that are intended to keep plans from becoming too underfunded. One such mechanism is the AFC, introduced by the Omnibus Budget Reconciliation Act of 1987 (OBRA '87). The AFC requires sponsors of plans with more than 100 participants that have become underfunded to a prescribed level to make additional plan contributions in order to prevent funding levels from falling too low. With some exceptions, plans with an actuarial value of assets below 90 percent of current liabilities are affected by the AFC rules. [Footnote 11] In addition to setting funding rules, ERISA established PBGC to guarantee the payment of the pension benefits of participants, subject to certain limits, in the event that the plan could not.[Footnote 12] Under ERISA, the termination of a single-employer DB plan may result in an insurance claim with the single-employer program if the plan has insufficient assets to pay all benefits accrued under the plan up to the date of plan termination.[Footnote 13] PBGC may pay only a portion of a participant's accrued benefit because ERISA places limits on the PBGC benefit guarantee. For example, PBGC generally does not guarantee benefits above a certain amount, currently $45,614 annually per participant at age 65.[Footnote 14] Additionally, benefit increases arising from plan amendments in the 5 years immediately preceding plan termination are not fully guaranteed, although PBGC will pay a portion of these increases.[Footnote 15] Further, PBGC's benefit guarantee amount is limited to the monthly straight life annuity benefit the participant would receive if she were to commence the annuity at the plan's normal retirement age.[Footnote 16] Sponsors of PBGC-insured DB plans pay annual premiums to PBGC for their coverage. Premiums have two components: a per participant charge paid by all sponsors (currently $19 per participant) and a "variable-rate" premium that some underfunded plans pay based on the level of unfunded benefits.[Footnote 17] The recent decline of PBGC's single-employer program has occurred in the context of the long-term stagnation of the DB system. The number of PBGC-insured plans has decreased steadily from approximately 110,000 in 1987 to about 29,000 in 2004. While the number of total participants in PBGC-insured single-employer plans has grown approximately 25 percent since 1980, the percentage of participants who are active workers has declined from 78 percent in 1980 to 50 percent in 2002. Unless something reverses these trends, PBGC may have a shrinking plan and participant base to support the program in the future. Many of the 100 Largest Plans' Liabilities Exceeded Plan Assets from 1995 to 2002, and Few Sponsors Were Required to Make Cash Contributions: From 1995 to 2002, while most of the 100 largest plans had sufficient assets to cover their plan liabilities, many did not. Furthermore, because of leeway in the actuarial methodology and assumptions sponsors can use to measure plan assets and liabilities, underfunding may actually have been more severe and widespread than reported at the end of the period. Because of flexible funding rules permitting the use of accounting credits other than cash contributions to satisfy minimum funding obligations, on average 62.5 of the 100 largest plans each year received no cash contributions from their sponsors. Although as a group, funding levels among the 100 largest plans were reasonably stable and strong from 1996 to 2000, by 2002, more than half of the largest plans were underfunded (see fig. 1). Two factors in the deterioration of many plans' finances were the decline in stock prices and prevailing interest rates. From 2000 to 2002, stock prices declined sharply each year, causing a decline in the value of many plans' pension assets. In addition, over the sample period, 30-year Treasury bond rates, which served as the benchmark for the rate used by plans to calculate pension liabilities, generally fell steadily, raising current liabilities. The combination of lower asset values and higher pension liabilities had a serious, adverse effect on overall DB plan funding levels. Figure 1: Almost One-Fourth of the Largest Pension Plans Were Less than 90 Percent Funded on a Current Liability Basis in 2002: [See PDF for image] [End of figure] Rules May Allow Reported Funding Levels to Overstate Current Funding Levels: Accurate measurement of a plan's liabilities and assets is central to the sponsor's ability to maintain assets sufficient to pay promised benefits, as well as to the transparency of a plan's financial health. Because many plans chose allowable actuarial assumptions and asset valuation methods that may have altered their reported liabilities and assets relative to market levels, it is possible that funding over our sample period was actually worse than reported for a number of reasons. These include the use of above-market rates to calculate current liabilities and actuarial measurement of plan assets that differ from market values. Reported current liabilities are calculated using a weighted average of rates from the 4-year period before the plan year. While this allows sponsors to smooth fluctuations in liabilities that sharp swings in interest rates would cause, thereby reducing volatility in minimum funding requirements, it also reduces the accuracy of liability measurement because the rate anchoring reported liabilities is likely to differ from current market values. To the extent that the smoothed rate used to calculate current liabilities exceeds current rates, the 4- year smoothing could reduce reported liabilities relative to those calculated at current market values. Further, rules allowed sponsors to measure liabilities using a rate above the 4-year weighted average.[Footnote 18] The 4-year weighted average of the reference 30- year Treasury bond rate exceeded the current market rate in 76 percent of time in the months between 1995 and 2002, and the highest allowable rate for calculating current liabilities exceeded the current rate in 98 percent of those months. Sponsors of the plans in our sample chose the highest allowable interest rate to value their current liabilities 62 percent of the time from 1995 to 2002. For example, an interest rate 1 percentage point higher than the statutorily required interest rate would decrease the reported value of a typical plan's current liability by around 10 percentage points. As with liabilities, the actuarial value of assets used for funding may also differ from current market values. Under the IRC, actuarial asset values cannot be consistently above or below market, but in a given year may be anywhere from 80 to 120 percent of market asset levels. Among the plans we examined, on average each year, 86 percent reported a different value for actuarial and market assets. On average, using the market value instead of the actuarial value of assets would have raised reported funding levels by 6.5 percent each year. However, while the market value exceeded the actuarial value of assets during the late 1990s, when plan funding was generally strong, in the weaker funding year of 2002 market assets dipped below actuarial assets. In 2001 and 2002, calculating plan funding levels using market assets would have greatly increased the number of plans below 90 percent funded each year. A similar calculation for 2002 would have drastically increased the number of large plans below 80 percent funded, from 6 to 24. Thus, we see some evidence that using actuarial asset values lowered the volatility of reported funding levels relative to those using market asset values. However, the use of the actuarial value of assets also may have disguised plans' funded status as their financial condition worsened. Two large plans that terminated in 2002 illustrate the potential effects of discrepancies between reported and actual funding. The Bethlehem Steel Corporation in 2002 reported that its plan was 85.2 percent funded on a current liability basis; yet, the plan terminated later that year with assets of less than half of the value of promised benefits. The PBGC single-employer program suffered a $3.7 billion loss as a result of that termination, its largest ever at the time. Similarly, LTV Steel Company reported that its pension plan for hourly employees was over 80 percent funded on its Form 5500 filing for plan year 2001. When this plan terminated in March, 2002, it had assets equal to 52 percent of benefits, a shortfall of $1.6 billion. Most Sponsors of Large Plans Did Not Make Annual Cash Contributions, but Satisfied Funding Requirements through Use of Accounting Credits: For the 1995 to 2002 period, the sponsors of the 100 largest plans each year on average made relatively small cash contributions to their plans. Annual cash contributions for the top 100 plans averaged approximately $97 million on plans averaging $5.3 billion in current liabilities, with figures in 2002 dollars.[Footnote 19] This average contribution level masks a large difference in contributions between 1995 and 2001, during which period annual contributions averaged $62 million, and in 2002, when contributions increased significantly to almost $395 million per plan. Further, in 6 of the 8 years in our sample, a majority of the largest plans made no cash contribution to their plan (see fig. 2). On average each year, 62.5 plans received no cash contribution, including an annual average of 41 plans that were less than 100 percent funded. Figure 2: Most Large Plans Received No Annual Cash Contribution, 1995- 2002: [See PDF for image] Note: Average contributions for 2002 are largely driven by one sponsor's contribution to its plan. Disregarding this $15.2 billion contribution reduces the average plan contribution for 2002 from $395 million to $246 million. [End of figure] The funding rules allow sponsors to meet their plans' funding obligations through means other than cash contributions. If a plan has sufficient FSA credits from other sources, such as an existing credit balance or large interest or amortization credits, to at least match its FSA charges, then the plan does not have to make a cash contribution in that year.[Footnote 20] Because meeting minimum funding requirements depends on reconciling total annual credits and charges, and not specifically on cash contributions, these other credits can substitute for cash contributions. From 1995 to 2002, it appears that many of the largest plan sponsors relied more heavily on other FSA credits than on cash contributions to meet minimum funding obligations. The average plan's credit balance carried over from a prior plan year totaled about $572 million (2002 dollars) each year, and 88 percent of plans on average carried forward a prior credit balance into the next plan year from 1995 to 2002. Not only could these accumulated credit balances help a plan to meet minimum funding obligations in future years, but they also accrue interest that augments a plan's FSA credits and further helps meet minimum funding requirements. In contrast, annual cash contributions averaged only $97 million, in 2002 dollars. On average each year, cash contributions represented 90 percent of the minimum required annual funding (from cash and credits).[Footnote 21] However, this average figure was elevated by high levels of contributions by some plans in 1995, 1996, and 2002. From 1997 to 2000, when funding levels were generally strong, cash contributions averaged only 42 percent of minimum required annual funding. During these years, a majority of plans in our sample received no cash contribution. Cash contributions represented a smaller percentage of annual minimum required funding during years when plans were generally well funded, indicating that in these years more plans relied more heavily on credits to meet minimum funding obligations. In addition to large credit balances brought forward from prior years, sponsors were able to apply funding credits from other sources, such as net interest credits ($42 million per plan per year, on average), and credits from the excess of a plan's calculated minimum funding obligation above the plan's full funding limitation ($47 million).[Footnote 22] Other plan events result in plan charges, which reflect events that increase the plan's obligations. For example, plans reported annual amortization losses, which could result from actual investment rates of return on plan assets below assumed rates of return (including outright losses) or increases in the generosity of plan benefits; these net amortization charges averaged almost $28 million in our sample. Funding credits, offset by charges, may help satisfy a plan's minimum funding obligation, substituting for cash contributions, and may explain why a significant number of sponsors made zero cash contributions to their plans in many years. The FSA credit accounting system provides some advantages to DB plan sponsors. Amortization rules require the sponsor to smooth certain events that affect plan finances over several years, and accumulated credit balances act as a buffer against swings in future funding requirements. These features often allow sponsors to better regulate their annual level of contributions, compared to annual fluctuations if funding were based strictly on yearly differences between the market value of plan assets and current liabilities. Similarly, current-law measurement and funding rules provide a plan with some ability to dampen volatility in required funding caused by economic events that may sharply change a plan's liabilities or assets. Pension experts told us that this predictability and flexibility make DB sponsorship more attractive to employers.[Footnote 23] However, the FSA accounting system, by smoothing annual contributions and liabilities, may distort a plan's funding level. For example, suppose a sponsor accrues a $1 million credit balance from making a contribution above the required minimum in a year. Suppose then that this $1 million purchases assets that lose all of their value by the following year. Even though the plan no longer had this $1 million in assets, the sponsor could still use that credit balance (plus interest on the credit balance) to reduce this year's contribution to the plan. Because of amortization rules, the sponsor would have to report only a portion of that lost $1 million in asset value as a plan charge the following year. Similarly, sponsors are required to amortize the financial effect of a change in a plan's benefit formula, which might result in increased benefits and therefore a higher funding obligation, over a 30-year period. Thus, even though higher benefits would immediately raise a plan's obligation to fund, the sponsor must spread this effect in the plan's FSA over 30 years. This disconnection between the reported and current market condition of plan finances raises the risk that plans will not react quickly enough to deteriorating plan conditions. Further, it reduces the transparency of plan financial information to stakeholders, such as participants, and investors. The experience of two large plans that terminated in a severely underfunded state help illustrate the potential disconnection between FSA accounting and the plan's true funded status. As stated earlier, the Bethlehem Steel Corporation and LTV Steel Company both had plans terminate in 2002, each with assets approximately equal to 50 percent of the value of benefits. Yet each plan was able to forgo a cash contribution each year from 2000 to 2002 by using credits to satisfy minimum funding obligations, primarily from large accumulated credit balances from prior years. Despite being severely underfunded, each plan reported an existing credit balance in 2002, the year of termination. Full Funding Limitation Rule May Have Allowed Some Plan Sponsors to Forgo Plan Contributions: Another possible explanation for the many instances in which sponsors made no annual cash contribution regards the full funding limitation (FFL). The FFL is a cap on minimum required contributions to plans that reach a certain funding level in a plan year.[Footnote 24] However, the FFL does not represent the contribution that would raise plan assets to the level of current liability. The FFL represents a "maximum minimum" contribution for a sponsor in a given year--a ceiling on the sponsor's minimum funding obligation for the plan. Between 1995 and 2002, rules permitted some plans with assets as low as 90 percent of current liability to reach the FFL, meaning that a plan could be considered fully funded without assets sufficient to cover all accrued benefits. The FFL is also distinct from the plan's annual maximum tax-deductible contribution. Because sponsors may be subject to an excise tax on contributions above the maximum deductible amount, the annual maximum contribution can act as a real constraint on cash contributions. Flexibility in the FFL rule has allowed many plan sponsors to take steps to minimize their contributions. In our sample, from 1995 to 2002 approximately two-thirds of the sponsors in each year made an annual plan contribution at least as large as the plan's FFL. However, in 65 percent of these instances, the sponsor had chosen the highest allowable rate to calculate current liability; using a lower rate to calculate current liability may have resulted in a higher FFL and, therefore, may have required a higher contribution. Further, the FFL was equal to zero for 60 percent of plans each year, on average. This means that these plans were permitted to forego cash contributions as a result of the FFL rule. This reflects the fact that if a plan's FFL equaled zero, that plan had assets at least equal to 90 percent of current liabilities that year and would not be required to make an additional contribution. The interaction between the FFL rule and the annual maximum tax- deductible contribution also has implications on the amount that plan sponsors can contribute. In some years, the maximum deductible contribution rules truly constrained some sponsors from making any cash contribution. In 1998, 50 of the 60 plans that contributed to the maximum deductible amount had a maximum deductible contribution of zero (see fig. 3). This meant that any cash contribution into those plans that year would generally subject the sponsor to an excise tax.[Footnote 25] For 37 of these plans, this was the case even if the sponsor had chosen the lowest statutorily allowed interest rate for plan funding purposes, which would have produced the highest calculated current liabilities. This constraint did not apply to as many plans in some other years. For example, in 1996, 52 plans contributed the maximum deductible amount. Thirty of these plans had a maximum deductible contribution of zero. Fourteen of the plans in this situation could not have made any additional contributions. However, the other 16 could have made at least some contributions by choosing a lower interest rate to raise their maximum deductible contribution level. Figure 3: For Selected Years from 1996 to 2002, Most Sponsors Contributed the Plan's Maximum Deductible Amount, Which for a Number of Plans Was Zero: [See PDF for image] Note: Years of analysis are not continuous, as the PBGC study on maximum deductible contributions was conducted for years shown. Information on maximum deductible contributions is missing for between 7 and 17 plans each year. Data for these plans were either missing or incomplete to calculate the plan contributions with respect to the maximum deductible contribution. [End of figure] Very Few Sponsors of Underfunded Large Plans Paid an AFC from 1995 to 2002: Funding rules dictate that a sponsor of a plan with more than 100 participants in which the plan's actuarial value of assets fall below 90 percent of liabilities, measured using the highest allowable interest rate, may be liable for an AFC in that year. More specifically, a plan that is between 80 and 90 percent funded is subject to an AFC unless the plan was at least 90 percent funded in at least 2 consecutive of the 3 previous plan years. A plan with assets below 80 percent of liabilities, calculated using the highest allowable rate, is assessed an AFC regardless of its funding history.[Footnote 26] Despite the statutory threshold of a 90 percent funding level for some plans to owe an AFC, in practice a plan needed to be much more poorly funded to become subject to an AFC. While about 10 plans in our sample each year had funding below 90 percent on a current liability basis, on average fewer than 3 plans each year owed an AFC (see fig. 4). From 1995 to 2002, only 6 of the 187 unique plans that composed the 100 largest plans each year were ever assessed an AFC,[Footnote 27] and these plans owed an AFC a total of 23 times in years in which they were among the 100 largest plans. By the time a sponsor owed an AFC, its plan had an average funding level of 75 percent, suggesting that by the time the AFC was triggered, the plan's financial condition was weak. Further, while we observed 60 instances between 1995 and 2002 in which a plan had funding levels between 80 and 90 percent, only 5 times was a plan in this funding range subject to an AFC. This would indicate that, in practice, 80 percent represented the realistic funding threshold for owing or avoiding the AFC. Figure 4: Most Plans Less Than 90 Percent Funded Were Not Assessed an AFC: [See PDF for image] [End of figure] Even with those plans subject to an AFC, other FSA credits may help a plan satisfy minimum funding obligations. Among plans in our sample assessed an AFC, the average annual AFC owed was $234 million, but annual contributions among this group averaged $186 million, with both figures in 2002 dollars. In addition, 61 percent of the time a plan was subject to an AFC, the sponsor used an existing credit balance to help satisfy its funding obligation. Over 30 percent of the time a plan was assessed an AFC, the funding rules allowed the sponsor to forgo a cash contribution altogether that year. Sponsors that owed an AFC had mixed success at improving their plans' financial conditions in subsequent years, and most of these plans remained significantly underfunded. Among the 6 plans that owed the AFC, funding levels rose slightly from an average 75 percent when the plan was first assessed an AFC to an average 76 percent, looking collectively at all subsequent years. All of these plans were assessed an AFC more than once. Again, terminated plans provide a stark illustration of weaknesses in the rules' ability to ensure sufficient funding. Bethlehem Steel's plan was assessed an AFC of $181 million in 2002, but the company made no cash contribution that year, just as it had not in 2000 or 2001, years in which the plan was not assessed an AFC. When the plan terminated in late 2002, its assets covered less than half of the $7 billion in promised benefits. LTV Steel, which terminated its pension plan for hourly employees in 2002 with assets of $1.6 billion below the value of benefits, had its plan assessed an AFC each year from 2000 to 2002, but for only $2 million, $73 million, and $79 million, or no more than 5 percent of the eventual funding shortfall. Despite these AFC assessments, LTV Steel made no cash contributions to this plan from 2000 to 2002. Both plans were able to apply existing credits instead of cash to fully satisfy minimum funding requirements. Large Plans' Sponsors' Credit Ratings Appear Related to Certain Funding Behavior and Represent Risk to PBGC: The recent funding experiences of large plans, especially those sponsored by financially weak firms, illustrate the limited effectiveness of certain current funding rules and represent a potentially large implicit financial risk to PBGC. The financial health of a plan sponsor may be key to plan funding decisions because sponsors must make funding and contribution decisions in the context of overall business operations. From 1995 to 2002, on average, 9 percent of the largest 100 plans were sponsored by a firm with a speculative grade credit rating, suggesting financial weakness and poor creditworthiness.[Footnote 28] Financial strength of plan sponsors' business operations has been a key determinant of risk to PBGC. Financially weak sponsors of large, underfunded plans are, by the nature of the insurance offered by PBGC, likely to cause the most financial burden to PBGC and other premium payers. For instance, PBGC typically trustees a plan when a covered sponsor is unable to financially support the plan, such as in the event of bankruptcy or insolvency. Current funding rules, coupled with the presence of PBGC insurance, may create incentives for financially distressed plan sponsors to avoid or postpone contributions and increase benefits. Many of the minimum funding rules are designed so that sponsors of ongoing plans may smooth contributions over a number of years. Sponsors that are in financial distress, however, may have a more limited time horizon and place other financial priorities above "funding up" their pension plans. To the extent that the presence of PBGC insurance causes financially troubled sponsors to alter their funding behavior, PBGC's potential exposure increases. Underfunded plans sponsored by financially weak firms pose the greatest immediate threat to PBGC's single-employer plans. PBGC's best estimate of the total underfunding of plans sponsored by companies with credit ratings below investment grade and classified by PBGC as "reasonably possible" to terminate was an estimated $96 billion as of September 30, 2004 (see fig. 5).[Footnote 29] Figure 5: Total Underfunding among All DB Plans, and among Those Considered by PBGC as Reasonably Possible for Termination, Has Increased Markedly since 2001: [See PDF for image] Note: Underfunding figures for non-reasonably possible plans represent the end of the calendar year, except for 2004, which represents the end of fiscal year 2004 (September 30, 2004). Figures for reasonably possible plans are taken as of the end of each fiscal year. [End of figure] Plans Sponsored by Financially Weak Firms Exhibit Riskier Funding Behavior: From 1995 to 2002, we observed that plans sponsored by speculative grade-rated firms had lower levels of average funding compared with the average for the 100 largest plans. For instance, the average funding of these plans was 12 percentage points lower on average than the funding level for all plans from 1995 to 2002. Plans sponsored by speculative grade-rated firms were also more likely to be underfunded. From 1995 to 2002, each year, on average, 18 percent of plans sponsored by speculative grade-rated firms had assets that were below 90 percent of current liability. Plans sponsored by nonspeculative grade-rated firms had just over half this incidence, or an average of 10 percent of plans funded below 90 percent of current liability. Large plans sponsored by firms with a speculative grade rating were also more likely to incur an AFC. While plans sponsored by speculative grade-rated firms accounted for only 9 percent of all plans that we examined over the 1995 to 2002 period, they accounted for just over one- third of all instances in which a sponsor was required to pay an AFC. In contrast, no high investment grade sponsors (those rated AAA or AA) were required to pay an AFC for this period. While the AFC is intended to be a backstop for underfunded plans, to the extent that plans sponsored by speculative grade-rated firms are considered to pose a significant risk for near-term termination, it may not be an effective mechanism for improving a plan's funding level. Plans sponsored by firms that are in financial distress are, by definition, having difficulty paying off debts and may be ill equipped to afford increased contributions to their plan. That is, the AFC itself may be a symptom of plan distress rather than a solution to improve a plan's funding level. Large plans with sponsors rated as speculative grade were also generally more likely to use the highest allowable interest rate to compute their current liability under the minimum funding rules. While a majority of sponsors from all credit rating categories used the highest allowable interest rate, over the entire 1995 to 2002 period, speculative grade-rated sponsors used the highest rate at an incidence 23 percentage points above the incidence for all other plans in the sample. The use of higher interest rates likely lowers a plan's reported current liability and minimum funding requirement. To the extent that this depresses cash contributions, such plans may have a higher chance of underfunding, thus creating additional financial risk to PBGC. PBGC's claims experience shows that financially weak plans have been a source of substantial claims. Of the 41 largest claims in PBGC history in which a rating was known, 39 of the plan sponsors involved were credit rated as speculative grade at least 3 years prior to termination (see fig. 6). These claims account for 67 percent of the value of total gross claims on the single-employer program from 1975 to 2004. Most of the plan sponsors involved in these claims were given speculative grade ratings for many more years prior to their eventual termination. Even 10 years prior to plan termination, 33 of these 41 claims involved sponsors rated as speculative grade. Figure 6: Over 80 Percent of Sponsors Associated with PBGC's Largest Termination Claims Had Speculative Grade Ratings 10 Years Prior to Termination: [See PDF for image] Note: Based on 41 of PBGC's largest gross claims in which the rating of the sponsor was known, representing over 67 percent of total gross claims from 1975 to 2004. These 41 claims may include sponsors with more than one plan and are not limited to those plans in our sample. Ratings based on S&P rating. [End of figure] Conclusions: Widely reported recent large plan terminations by bankrupt sponsors and the financial consequences for PBGC have pushed pension reform into the spotlight of national concern. Our analysis here suggests that certain aspects of the funding rules have contributed to the general underfunding of pensions and, indirectly, to PBGC's recent financial difficulties. The persistence of a large number of underfunded plans, even during the strong economic period of the late 1990s, implies that current funding rules are not stringent enough to ensure that sponsors can fund their pensions adequately. Further, the rules appear to lack strong mechanisms to compel sponsors to make regular contributions to their plans, even those that are underfunded or subject to an AFC. Perhaps most troubling is that current rules for measuring and reporting plan assets and liabilities may not reflect true current values and often understate the true degree of underfunding. The current rules have the reasonable and important goals of long-term funding adequacy and short-term funding flexibility. However, our work shows that although the current system permits flexibility, it also permits reported plan funding to be inadequate, misleading, and opaque; even so, funding and contributions for some plans can still swing wildly from year to year. This would appear not to serve the interest of any DB pension stakeholders effectively. The challenge is determining how to achieve a balance of interests: how to temper the need for funding flexibility with accurate measurement, adequate funding, and appropriate transparency. Despite flaws in the funding rules, our work here shows that most of the largest plans appear to be adequately funded. Rules should acknowledge that funding will vary with cyclical economic conditions, and even sponsors who make regular contributions may find their plans underfunded on occasion. Periodic and mild underfunding is not usually a major concern, but it becomes a threat to workers' and retirees' economic security in retirement and to PBGC when the sponsor becomes financially weak and the risk of bankruptcy and plan termination becomes likely. This suggests that perhaps the stringency of certain funding rules should be adjusted depending on the financial strength of the sponsor, with stronger sponsors being allowed greater latitude in funding and contributions than weaker sponsors that might present a near-term bankruptcy risk. However, focusing more stringent funding obligations on weak plans and sponsors alone may not be adequate, because strong companies and industries can quickly become risky ones, and, once sponsors and plans become too weak, it may be difficult for them to make larger contributions and still recover. It should be noted also that while funding rule change is an essential piece of the overall reform puzzle, it is certainly not the only piece. Indeed, pension reform is a challenge precisely because of the necessity of fusing together so many complex, and sometimes competing, elements into a comprehensive proposal. Ideally, effective reform would: * improve the accuracy of plan asset and liability measurement while minimizing complexity and maintaining contribution flexibility; * develop a PBGC insurance premium structure that charges sponsors fairly, based on the risk their plans pose to PBGC, and provides incentives for sponsors to fund plans adequately; * address the issue of severely underfunded plans making lump-sum distributions; * resolve outstanding controversies concerning cash balance and other hybrid plans by safeguarding the benefits of workers regardless of age; and: * improve plan information transparency for PBGC, plan participants, unions, and investors in a manner that does not add considerable burden to plan sponsors. As deliberations on reform move forward, it will be important that each of these individual elements be designed so that all work in concert toward well-defined goals. Even with meaningful, carefully crafted reform, it is possible that some DB plan sponsors may choose to freeze or terminate their plans. While these are serious concerns, the overarching goals of balanced pension reform should be to protect the retirement benefits of American workers and retirees by providing employers reasonable funding flexibility while also holding those employers accountable for the promises they make to their employees. As I noted in my opening remarks, PBGC's challenges parallel the challenges facing our Social Security system. While both programs have adequate current revenues and assets to pay promised benefits today, both face large and growing accumulated deficits on an accrual basis. Further, timely action to address both private pension and Social Security reform is needed. However, consideration must be given to the interactive effects of any such reforms and how they contribute to addressing our nation's large and growing fiscal challenge, key demographic, economic and workforce trends, and the economic security of Americans in their retirement years. Mr. Chairman, this concludes my statement. I would be happy to respond to any questions you or other Members of the Committee may have. Contact and Acknowledgments: For further information, please contact Barbara Bovbjerg at (202) 512- 7215. Contact points for our Office of Congressional Relations and Public Affairs may be found on the last page of this testimony. Other individuals making key contributions to this testimony included Charlie Jeszeck, Mark Glickman, and Chuck Ford. FOOTNOTES  See GAO, Private Pensions: Recent Experiences of Large Defined Benefit Plans Illustrate Weaknesses in Funding Rules, GAO-05-294 (Washington, D.C.: May 31, 2005).  See GAO, Pension Benefit Guaranty Corporation: Single-Employer Pension Insurance Program Faces Significant Long-Term Risks, GAO-04-90 (Washington, D.C.: Oct. 29, 2003).  See GAO, 21st Century Challenges: Re-Examining the Base of the Federal Government, GAO-05-325SP (Washington, D.C.: Feb. 2005).  We analyzed DB pension data for the 100 largest plans as ranked by current liabilities reported on Schedule B of the Form 5500 for the years 1995 to 2002. The Form 5500 is a disclosure form that private sector employers with qualified pension plans are required to file with the Internal Revenue Service (IRS), Labor's Employee Benefit Security Administration (EBSA), and PBGC. While our sample of plans represented only a small portion of the total plans in the single-employer program, it constitutes approximately 50 percent of the total liabilities and about 28 percent of the total participants among DB plans that filed a Form 5500 in 2002. For more information on our methodology, see appendix I of GAO-05-294.  An underfunded plan does not necessarily indicate that the sponsor is unable to pay current benefits. Underfunding means that the plan does not currently have enough assets to pay all accrued benefits, the majority of which will be paid in the future, under the given actuarial assumptions about asset rate of return, retirement age, mortality, and other factors that affect the amount and timing of benefits.  Present value calculations reflect the time value of money--that a dollar in the future is worth less than a dollar today, because the dollar today can be invested and earn interest. Using a higher interest rate will lower the present value of a stream of payments because it implies that a lower level of assets today will be able to fund those future payments.  Normal cost is the cost of pension benefits allocated to a specific plan year.  Plans may amortize experience gains or losses over a 5-year period. Changes in the terms of the plan arising from plan amendments may be amortized over a 30-year period. Thus, these events continue to affect the FSA and plan funding for several years after they occur.  The rate used to calculate current liability has usually been based on the 30-year Treasury bond rate, with the allowable range above and below the 4-year weighted average varying in different years. The Pension Funding Equity Act of 2004 replaced the Treasury bond rate with the corporate index for plan years 2004 and 2005. See IRC section 412(b)(5)(B)(ii)(II). For further discussion of rates used to discount pension liabilities, see GAO, Private Pensions: Process Needed to Monitor the Mandated Interest Rate for Pension Calculations, GAO-03-313 (Washington, D.C.: Feb. 27, 2003).  A plan's current liability may differ from its "termination liability," which measures the value of accrued benefits using assumptions appropriate for a terminating plan. For further discussion of current versus termination liability, see GAO-04-90, appendix IV.  A single-employer plan may be subject to an AFC in a plan year if plan assets fall below 90 percent of current liabilities. However, a plan is not subject to an AFC if the value of plan assets (1) is at least 80 percent of current liability and (2) was at least 90 percent of current liability for at least 2 consecutive of the 3 immediately preceding years. To determine whether the AFC applies, the IRC requires sponsors to calculate current liabilities using the highest interest rate allowable for the plan year. See 26 U.S.C. 412(l)(9)(C).  Some DB plans are not covered by PBGC insurance; for example, plans sponsored by professional service employers, such as physicians and lawyers, with 25 or fewer active participants.  The termination of a fully funded DB plan is called a standard termination. Plan sponsors may terminate fully funded plans by purchasing a group annuity contract from an insurance company, under which the insurance company agrees to pay all accrued benefits, or by paying lump-sum benefits to participants if permissible. The termination of an underfunded plan, termed a distress termination, is allowed if the plan sponsor requests the termination and the sponsor satisfies other criteria. Alternatively, PBGC may initiate an "involuntary" termination. PBGC may institute proceedings to terminate a plan if the plan has not met the minimum funding standard, the plan will be unable to pay benefits when due, a reportable event has occurred, or the possible long-run loss to PBGC with respect to the plan may reasonably be expected to increase unreasonably if the plan is not terminated. See 29 U.S.C. 1342(a).  This guarantee level applies to plans that terminate in 2005. The amount guaranteed is adjusted (1) actuarially for the participant's age when PBGC first begins paying benefits and (2) if benefits are not paid as a single-life annuity. Because of the way ERISA allocates plan assets to participants, certain participants can receive more than the PBGC guaranteed amount.  The guaranteed amount of the benefit amendment is calculated by multiplying the number of years the benefit increase has been in effect, not to exceed 5 years, by the greater of (1) 20 percent of the monthly benefit increase calculated in accordance with PBGC regulations or (2) $20 per month. See 29 C.F.R. 4022.25(b).  For more on PBGC guarantee limits, see Pension Benefit Guaranty Corporation, Pension Insurance Data Book 1999 (Washington, D.C., Summer 2000), pp. 2-14.  The additional premium equals $9.00 for each $1,000 (or fraction thereof) of unfunded vested benefits. However, no such premium is charged for any plan year if, as of the close of the preceding plan year, contributions to the plan for the preceding plan year were not less than the full funding limitation for the preceding plan year.  In 1987, the permissible range was not more than 10 percent above, and not more than 10 percent below, the weighted average of the rates of interest on 30-year Treasury bond securities during the 4-year period ending on the last day before the beginning of the plan year. The top of the permissible range was gradually reduced by 1 percent per year, beginning with the 1995 plan year, to not more than 5 percent above the weighted average rate effective for plan years beginning in 1999. The top of the permissible range was increased to 20 percent above the weighted average rate for 2002 and 2003. For 2004 and 2005, the Congress changed the reference rate from the 30-year Treasury bond rate to a rate based on long-term investment-grade corporate bonds, and reset the allowable range for plans to 90 to 100 percent of this rate.  For the 100 largest plans that we examined, all dollar figures are reported in constant 2002 dollars.  If FSA credits exceed charges in a given plan year, the plan's FSA registers a net "credit balance" that may be carried forward to the next plan year; conversely, a prior year's funding deficiency also carries forward. The FSA credit balance at year-end is equal to the FSA credit balance at the beginning of the year plus FSA credits less FSA charges. Compliance with the minimum funding standard requires that the FSA balance at the end of the year is non-negative. An existing credit balance accrues interest and may be drawn upon to help satisfy minimum funding requirements for future plan years, and it, therefore may offset the need for future cash contributions.  Minimum required annual funding equals annual total FSA charges, less net amortization credits and interest applied to these amortization credits.  Full funding limitation rules set a ceiling for minimum annual funding requirements for a plan each year, based on the plan's liabilities.  There are investment techniques, such as purchasing fixed income assets whose payouts match the plan's expected payouts, which could make pension funding relatively predictable, even without FSA smoothing. One possible reason that such techniques are not widely used may be they are believed to be more expensive, over the long term than an asset allocation with significant equity investment exposure.  As with other funding rules, determining a plan's FFL is complicated. From 1995 to 2002, the FFL equaled the higher of (1) 90 percent of the plan's current liability or (2) the lower of (a) the accrued plan liability or (b) 150 to 170 percent (depending on the year) of the current liability. As of the 2004 plan year, the 150 to 170 percent measure no longer factors in the determination of the FFL. For our sample of plans, an average of 4 plans per year were above 150 to 170 percent (depending on the year) of the current liability and had an FFL of zero. This means the sponsors of these plans were most likely unable to make additional contributions unless they paid an excise tax.  For years after 2001, an employer may elect not to count contributions as nondeductible up to the full-funding limitation that is based on the accrued liability. Therefore, it could be possible for a sponsor to contribute more than the maximum deductible amount and still avoid the excise tax. See 26 U.S.C. 4972(c)(7).  The rules for determining the amount of the AFC are complex, but they generally call for sponsors to pay a percentage of their unfunded liability. Under current law, plans that owe an AFC may still apply FSA credits to meet their funding obligation and, therefore, may not be required to satisfy the AFC with a cash contribution.  Unique plans refer to the number of plans we observed with distinct plan identifiers called EINs and PINs. See footnote 9 of GAO- 05-294 for further information on why the actual number of completely unrelated plans in our sample may be lower than the 187 reported.  Credit ratings are generally considered to be a useful proxy for a firm's financial health. A credit rating, generally speaking, is a rating service's current opinion of the creditworthiness of an obligor with respect to a financial obligation. It typically takes into consideration the creditworthiness of guarantors, insurers, or other forms of credit enhancement on the obligation and takes into account the currency in which the obligation is denominated. Moody's and Standard and Poor's (S&P) are two examples of well-known ratings services.  Criteria used for classifying a plan as a reasonably possible termination include, but are not limited to, one or more of the following conditions: the plan sponsor is in Chapter 11 reorganization; funding waiver pending or outstanding with the Internal Revenue Service; sponsor missed minimum funding contribution; sponsor's bond rating is below-investment-grade for Standard & Poor's (BB+) or Moody's (Ba1); sponsor has no bond rating but unsecured debt is below investment grade; or sponsor has no bond rating, but the ratio of long- term debt plus unfunded benefit liability to market value of shares is 1.5 or greater.