FROM THE OFFICE OF PUBLIC AFFAIRS
October 14, 2003
JS-919
Testimony of Mark J. Warshawsky
Acting Assistant Secretary for Economic Policy
U.S. Department of the Treasury
Before the Special Committee on Aging
United States Senate
The Administration’s Activities To Improve The Retirement Security Of Defined Benefit Pension Participants
Chairman Craig, Ranking Member Breaux, and distinguished members of the Committee. I am pleased to appear before you with Pension Benefit Guaranty Corporation (PBGC) Executive Director Steven Kandarian and William Sweetnam, Benefits Tax Counsel of the U.S. Treasury, to discuss defined benefit pension plans. I will discuss the Administration’s proposals and ongoing activities aimed at strengthening the long-term health of the defined benefit pension system and thereby improving the retirement security of defined benefit pension participants. Bill and I will be happy to answer any questions you may have.
We all want to improve the retirement security for the nation’s workers and retirees by strengthening the financial health of the voluntary defined benefit system that they rely upon. We believe that with improvements, the defined benefit system will continue to be a viable and important part of the American retirement system. Despite repeated attempts to improve the current defined benefit pension funding system, it seems that, even without the impact of market downturns over the past few years, conditions have worsened over time. PBGC's current estimate suggests that pension plans in aggregate are underfunded by more than $350 billion. PBGC’s most recent unaudited figures show liabilities outstripping assets by $8.8 billion.
Before discussing comprehensive reform, I would like to discuss the proposals that the Administration has already put forward in this area. In July, we released the Administration’s Proposal to Improve the Accuracy and Transparency of Pension Information. This proposal is designed to strengthen and secure Americans’ pension security by:
- Improving the accuracy of the pension liability discount rate;
- Increasing the transparency of pension plan information; and
- Strengthening safeguards against pension underfunding.
THE ADMINISTRATION’S PROPOSAL FOR ACCURATELY MEASURING PENSION LIABILITIES
Fixing the pension funding rules won’t help unless we give our immediate attention to ensure that we accurately measure the pension liabilities on which those rules rely. Our most immediate task is replacing the 30-year Treasury rate used in measuring pension liabilities for minimum funding purposes. The Administration’s proposal is the necessary first step in the reform process. The Administration believes that any permanent change in pension discounting rules should not contribute to future pension plan underfunding. The Administration seeks to have pension liabilities accurately measured, in order to provide the necessary foundation for reform of the funding rules. Once we know the extent of these pension liabilities, we can ensure that pension promises made are pension promises kept.
We appreciate that there is important activity in both Houses of Congress on the issue. In the Senate, the Finance Committee Chairman’s Modification to the “National Employee Savings and Trust Equity Guarantee Act of 2003” includes a discounting provision that is quite similar to the Administration’s proposal. We were happy to see that provision included in the bill.
On the House side, the Administration believes that H.R. 3108, the Pension Funding Equity Act of 2003, is an important first step toward providing a permanent replacement for the interest rate now used to determine pension liabilities. H.R. 3108’s proposed discounting method for the next two years is broadly consistent with the Administration’s proposal over the same time frame. We are encouraged by the passage of this bill.
We face two near-term concerns that must be addressed in getting to a permanent replacement of the current discount rate.
First, firms that sponsor defined benefit plans already are budgeting their pension contributions for the next several years. Near-term changes to the current rules that would increase pension contributions above current expectations could disrupt these firms’ existing short-term plans.
Second, many underfunded plans are already facing sharp increases in their required pension funding contributions. Thus, while we must ultimately ensure that liabilities are measured accurately and that firms appropriately fund the pension promises they have made, an abrupt change from the current system could do more short-term harm than good by triggering plan freezes or terminations.
The Importance of the Discount Rate in Pension Funding
To determine minimum required funding contributions, a plan sponsor must compute the present value of the plan participants’ accrued future benefit payments, which is known as the plan’s current liability. The present value of a benefit payment due during a particular future year is calculated by applying a discount factor to the dollar amount of that payment. This discount factor converts the dollar value of the future payment to today’s dollars. Current liability is simply the sum of all these discounted future payments.
Pension liabilities must be accurately measured to ensure that pension plans are adequately funded to protect workers’ and retirees’ benefits and to ensure that minimum funding rules do not impose unnecessary financial burdens on plan sponsors. Liability estimates that are too low will lead to plan underfunding, potentially undermining benefit security. Pension plan liability estimates that are too high lead to higher than necessary minimum contributions, reducing the likelihood that sponsors will continue to operate defined benefit plans.
Computing pension liabilities is basically a two-step process. In the first step, the plan actuary estimates the payments that will be made to retirees each year in the future. The pension plan’s actuary makes these estimates based on the plan’s terms, and estimates of how long current employees will work before retirement and receive benefits in retirement. Estimating the future stream of payments involves considerable judgment on the part of the actuary.
Step two, converting the value of future payments to today’s dollars, is, by comparison, simple and rather mechanical. To convert payments in a future year to present dollars, the estimated payments are simply adjusted by the appropriate discount rate. Although some discounting schemes use the same discount rate to compute the present value of payments for all future years, it is no more difficult to compute the present value using different discount rates for each future year.
Choosing the right rate is the key to accurate pension discounting. The wrong rate leads to inaccurate estimates of liabilities that can be either too high or too low.
Therefore, the primary goal of the Administration’s proposal to replace the 30-year Treasury rate can be summed up in one word: accuracy. Without first accurately measuring a plan’s pension liabilities, the minimum funding rules cannot ensure that the firm is setting aside sufficient funds to make good on its pension promises to its workers. Accurate liability measures also provide a firm’s investors with valuable information about the pension contributions that will be made from the firm’s earnings. Accurate liability measures allow workers and retirees to monitor the health of their pension plans. Finally, accurate liability measures allow the PBGC to better monitor the health of the overall pension system.
Pension Discounting under Current Law
Since 1987, federal law has required that pension liabilities that determine minimum pension contributions be computed using the interest rate on the 30-year Treasury bond. In 2002, Congress passed legislation that temporarily changed the discount rate to provide funding relief to plan sponsors. This temporary fix expires at the end of this year.
Dissatisfaction with the continued use of the 30-year rate, even on an interim basis, has been expressed by many Members of Congress and pension sponsors. This dissatisfaction and the recognition that the 30-year rate is no longer an accurate discount rate make it imperative that a replacement be promptly enacted. This is why the Administration applauds the passage of the House Bill.
The Administration’s Proposal for Accurately Measuring Pension Liabilities
The Administration believes that corporate bond rates, not Treasury rates, should be the basis for the pension discount methodology. Three key issues need to be addressed in selecting a permanent replacement for the 30-year Treasury rate: the time structure of a pension plan’s future benefit payments; the appropriateness of smoothing the discount rate; and the appropriate relationship between the discount rate and the computation of lump sum payments.
The proposal I will now set forth deals with each of these issues.
1. Pension discount rates should be based on market determined interest rates for similar obligations.
The terms of pension contracts are not market determined because pensions are not bought and sold in an open market and pension sponsors do not compete with one another for participants. However, group annuity contracts, which are very similar to employer sponsored pensions, are sold in a competitive market by insurance companies. Group annuity contracts obligate the seller to provide a stream of annual cash payments, in exchange for a competitively priced premium, to individuals covered by the policy. We take the view, as Congress has in the past, that pension discount rates should reflect the interest rate underlying group annuity prices. These assets held by annuity providers consist largely of bonds issued by firms with high credit ratings. Furthermore, the insurance companies issuing the group annuity contracts also have high credit ratings.
Therefore, the Administration proposes that the new pension discount rate be based upon an index of interest rates on high-grade corporate bonds.
2. Pension discount rates should be designed to ensure that liabilities reflect the timing of future benefit payments.
Each pension plan has a unique schedule of future benefit payments - or cash flow profile - that depends on the characteristics of the work force covered by the plan. These characteristics include the percent of participants that are retired, the age of current workers covered by the plan, the percent receiving lump sums and whether the covered work force has been growing or shrinking over time. Plans with more retirees and older workers, more lump sum payments, and shrinking workforces will make a higher percentage of their pension payments in the near future, while plans with younger workers, fewer retirees, fewer lump sums, and growing workforces will make a higher percentage of payments in later years.
One approach to liability computation applies the same discount rate to all future payments regardless of when they occur. This approach produces inaccurate liability estimates because it ignores a basic reality of financial markets: that the rate of interest earned on an investment or paid on a loan varies with the length of time of the investment on the loan. If a consumer goes to a bank to buy a Certificate of Deposit, he will expect to receive a higher rate on a five-year CD than on a one-year CD. Likewise, that same consumer who borrows money to buy a house expects to pay a higher interest rate for a 30-year than a 15-year mortgage.
Pension discount rates must recognize this simple financial reality. Pension payments due next year should be discounted at a different, and typically lower, rate than payments due 20 years from now. Why is this important? Pension plans covering mostly retired workers that use a 20-year interest rate to discount all their benefit payments will understate their true liabilities. This will lead to plan underfunding that could undermine retiree pension security, especially for workers who are nearing retirement age. Proper matching of interest rates to payment schedules cannot be accomplished using any single discount rate.
Computing liabilities by matching interest rates on zero-coupon bonds that mature on the same date that benefit payments are due is not complicated. Once expected pension cash flows are calculated by the actuary it is no more difficult to discount benefit payments on a spreadsheet with an array of different interest rates than it is if only one discount rate is used.
It is also important to understand that the discount rate used does not change the actual obligation -- the liability is what it is. Choosing the proper discount rate gives us an accurate measure in today’s dollars of future benefit payments; it does not change those payments. But if we don’t measure that value properly today, plans may not have sufficient funds set aside in the future to make good on those pension promises.
The Administration proposes that benefit payments made in future years be discounted to today’s dollars using discount rates taken from a corporate bond yield curve (a table or graph that illustrates the interest rates on bonds that mature at different dates in the future). Liabilities would be computed by using interest rates on zero-coupon bonds that mature on a specific date in the future to discount benefit payments due to be made that same year.
Furthermore, implementation of the yield curve would be phased in over five years. The phase-in would start with the use of a single long-term corporate bond rate for the first two years. In the third year a phase-in to the appropriate yield curve discount rate would begin. The yield curve would be fully applicable by the fifth year.1
This phase-in period would provide some short term funding relief for sponsors, and achieve the desired level of accuracy at the end of five years.
3. Pension discount rates should be based on current financial conditions.
Pension liability computations should reflect the current market value of future benefit payments -- this is a key component of accuracy. Plan sponsors and investors are interested in the current value of liabilities in order to determine the demands pension liabilities will place on the company’s future earnings. Workers and retirees are interested in the current value of liabilities so that they can determine whether their plans are adequately funded.
Some argue that discount rates should be averaged (smoothed) over long periods of time. Under current law they are smoothed over four years. Such smoothing is intended to reduce the volatility of liability measures and helps make contribution requirements more predictable. Unfortunately, current smoothing rules reduce the accuracy of liability measures while failing to achieve stability in annual contributions. Smoothing can mask changes in pension plan solvency of which workers and retirees should be aware. As I mentioned earlier, we would like to work with Congress to identify permanent reforms of the funding rules that would reduce volatility in annual contributions, without the corollary effect of reducing measurement accuracy.
The Administration proposes to decrease smoothing gradually during the five-year phase-in. In years one and two, four year smoothing is maintained. Smoothing is reduced in years three and four and finally, in year five, set at a 90-day moving average to eliminate the impact of day-to-day market volatility. This will provide an appropriately current measure of interest rates.
4. Pension discount rates should apply to annuities and lump sum payments in a consistent and neutral manner.
Retirees and departing workers in some plans can opt to receive a single payment for their pension benefits rather than regular payments over their lifetimes. The value of these so-called lump sum payments is the present value of the worker’s expected retirement annuity. Using an artificially low discount rate for lump sums creates an incentive for participants to choose lump sums rather than the annuity
The Administration proposes that the yield curve used to measure pension liabilities also be used to compute lump sum payments so as to reflect accurately the life expectancy of retirees in the amounts that they will receive. In order to minimize the disruption of plans of workers who will receive benefits in the immediate future, lump sums would be computed using the 30-year Treasury rate as under current law in years one and two. In the third year a phase-in to the appropriate yield curve discount rate would begin. By the fifth year lump sums will be computed using the yield curve.
Workers receiving lump sums, especially those in their 50's, 60's and older, would be better off under the Administration’s proposal than under an alternative that would compute lump sums using a single long term corporate interest rate. Workers electing lump sums at relatively younger ages would have a higher proportion of their future payments discounted at long-term interest rates than workers retiring at relatively older ages. This is appropriate given the different time frames over which they had been expecting to receive their benefits. While moving from the 30-year Treasury rate to any corporate bond based rate will result in lower lump sum payments for younger workers who leave their jobs, under the yield curve approach older workers closer to retirement age will be little affected by the change.
However, some workers who will soon be leaving their jobs have been anticipating taking their pension benefits in the form of a lump sum with the expectation that those benefits would be computed using the 30-year Treasury rate. Computing lump sums using the yield curve rather than the 30-year Treasury rate may result in lower lump sum payments for those who leave at a young age. The Administration’s proposal is for the benefits of younger and older workers alike to be consistently and accurately valued, whether a lump sum or a traditional annuity benefit.
Development and Use of the Yield Curve
Yield curves used to discount pension benefit payments have been available for a number of years. One example of such a pension yield curve is the one developed by Salomon Brothers (now Citibank) in 1994 for the Securities and Exchange Commission. Monthly Citibank yield curves can be found on the Society of Actuaries web site at www.soa.org. In the past months, the Treasury Department has begun active development of our own yield curve based on interest rates for high-quality, zero-coupon, call adjusted corporate bonds of varying maturities using an alternative widely accepted methodology. We are very pleased with our progress in this regard and do not foresee any difficulty in generating yield curves for use in discounting pension plan payments if the Administration’s proposal becomes law.
Treasury would use a formal notice and comment rulemaking process to ensure transparency and to incorporate input from all interested parties in final development of the yield curve. Although the groundwork is well established, we certainly plan to work with all stakeholders to finalize the methodological details of the ultimate yield curve.
Because discounting pension payments using a yield curve is already considered a best practice in financial accounting, large sponsors are almost certainly making these computations now or know how to make them.2 Sponsors certainly know what their expected future pension cash flows are.
The mechanics of discounting future pension cash flows are in fact quite simple. This is true whether one uses a single rate to discount all payments or uses different rates to discount payments made in each year. Such calculations, which can be done with a simple spreadsheet, should not pose serious problems even for small plans let alone plans sponsored by large, financially sophisticated firms.
As I stated at the outset, the Administration’s permanent discount rate replacement proposal is designed to strengthen American’s retirement security by producing accurate measures of pension liabilities. And accurate measurement is the essential first step in ensuring that pension promises made are pension promises kept.
ADMINISTRATION PROPOSALS TO INCREASE TRANSPARENCY AND STRENGTHEN PENSION FUNDING
There are two other reform tasks that the Administration recommends for immediate attention. First, the transparency of information pertaining to pension plan funding needs to be increased. Under current law most workers and retirees are not provided with timely information about the funding of their pension plans. We propose to remedy this by requiring that each year sponsors disclose to participants the value of their defined benefit pension plan assets and liabilities measured on both a current liability and a termination liability basis.
The Administration also proposes that certain financial data already collected by the PBGC from companies sponsoring pension plans with more than $50 million of underfunding should be made public. We propose that the available information be limited to the underfunded plan's market value of assets, termination liability and termination funding ratios. This data is more timely and accurate than what is publicly available under current law.
Second, the Administration proposes to restrict benefit increases for certain underfunded plans whose sponsors are financially troubled. When firms with below investment grade credit ratings increase pension benefit promises, the costs of these added benefits stand a good chance of being passed on to the pension insurance system, frustrating the benefit expectations of workers and retirees and penalizing employers who have adequately funded their plans. Under the Administration's proposal, if a plan sponsored by a firm with a below investment grade credit rating has a funding ratio below 50 percent of termination liability, benefit improvements would be prohibited, the plan would be frozen (no accruals resulting from additional service, age or salary growth), and lump sum payments would be prohibited unless the employer contributes cash or provides security to fully fund these added benefits. When a plan sponsor files for bankruptcy the PBGC’s guarantee limits would also be frozen.
It should also be noted that Treasury is in the process of updating mortality assumptions. In order to ensure that liabilities are measured accurately, mortality estimates need to be made from the most up-to-date and accurate tables available. On September 22, 2003, the Treasury and the Internal Revenue Service published in the Internal Revenue Bulletin, a request for comments on the mortality tables used in determining current liabilities. The notice invites comments on methods of projecting mortality and on factors, in addition to age and year of birth, that might be appropriately reflected in any new tables that may be adopted.
FUNDAMENTAL REFORM
Currently both the Senate and House bills contain calls for comprehensive reform. The Administration supports and appreciates these provisions and looks forward to working with Congress on this important issue. The Administration commends those in Congress who have recognized that there is a need for reform and we look forward to your continued leadership on these issues. Americans have a broadly shared interest in adequate funding of employer-provided defined benefit pensions. Without adequate funding, the retirement income of America’s workers will be insecure. This by itself is a powerful reason to pursue improvements in our pension system. At the same time, we must always be mindful that the defined benefit pension system is voluntary. Firms offer defined benefit pensions to their workers as an employee benefit, as a form of compensation. Our pension rules should thus be structured in ways that encourage, rather than discourage, employer participation.
Key aspects of the current system frustrate participating employers while also failing to produce adequate funding. We thus have multiple incentives to improve our pension system, and to thus better ensure both the availability and the viability of worker pensions. We have rolled up our sleeves and begun the hard work needed to create a system that more clearly and effectively funds pension benefits. We will develop a pension system that will be less complex, more flexible, logically consistent, and will achieve the goal of improving the security of defined benefit plans. Major areas that require our prompt attention include:
1. Funding Targets
We will seek to develop better, more meaningful, funding targets.
Asset Measurement. Under existing rules, assets can be measured as multi-year averages rather than current values. Pension funding levels can only be set appropriately if both asset and liability measures are current and accurate. Failure to accurately measure assets and liabilities contributes to funding volatility.
Liability Measurement
We also intend to examine how the application of actuarial assumptions in the current rules may contribute to funding volatility and to inaccurate measurement of pension liabilities. For example, companies do not want to be surprised to find they have inadequately funded their plans because the mortality tables used in the funding rules are outdated or because those rules fail to account for lump sum payments. We will examine:
a. Retirement Assumptions. Retirement assumptions made by plan actuaries need to reflect the actual retirement behavior of those covered by the plan.
b. Lump Sums. Liability computations for minimum funding purposes need to include reasonable estimates of expected future lump sum withdrawals that are determined by methodologies that are broadly consistent with other estimates of plan obligations.
c. Mortality. As noted above, Treasury is in the process of updating mortality assumptions.
2. Funding Path
The current system of funding rules and asset and liability measurement has been constructed, in part, to dampen the volatility of firms’ funding contributions. Yet current rules fail to do so. After years of making few or no contributions at all, many firms are facing precipitous increases in their annual funding requirements. This outcome is frustrating to business and it has failed to provide adequate funding for workers and retirees. Improvements to funding rules should mitigate volatility, provide firms with the ability to make more consistent contributions, and increase flexibility for firms to fund up their plans in good times. Specific issues in the funding rules that need to be examined include:
a. Contribution Deductibility. Together, minimum funding rules and limits on maximum deductible contributions require sponsors to manage their funds within a narrow range. Raising the limits on deductible contributions would allow sponsors to build larger surpluses to provide a better cushion for bad times.
b. Credit Balances. If a sponsor makes a contribution in any given year that exceeds the minimum required contribution, the excess plus interest can be credited against future required contributions. These credit balances - mere accounting entries - do not fall in value even if the assets that back them lose value. Credit balances allow seriously underfunded plans to avoid making contributions, often for years, and contribute to funding volatility.
c. Volatility Caused by the Minimum Funding Backstop. The current minimum funding backstop, known as the deficit reduction contribution, causes minimum contributions of underfunded plans to be excessively volatile from year to year.
d. New Benefit Restrictions. The current Administration proposal is to restrict benefit increases for certain underfunded plans whose sponsors are financially troubled. We are looking at areas where it may be appropriate to expand this proposal.
e. Benefit Amortization. The amortization period for new benefits can be up to 30 years long. This may be excessive. We will also look at other statutorily defined amortization periods.
3. Other Issues
a. Extent of Benefit Coverage. It may be advisable to limit or eliminate guarantees of certain benefits that typically are not funded, such as shutdown benefits.
b. Multi-employer Plan Problems. Multi-employer plans operate under a different set of rules than single-employer plans. Despite these regulatory differences, the same principles of accuracy and transparency should apply to multi-employer plans, and we will be reviewing the best ways to accomplish this.
c. PBGC Premiums. PBGC’s premium structure should be re-examined to see whether it can better reflect the risk posed by various plans to the pension system as a whole.
CONCLUSION
As I stated at the outset, the Administration’s permanent discount rate replacement proposal is designed to strengthen American’s retirement security by producing accurate measures of pension liabilities. And accurate measurement is the essential first step in ensuring that pension promises made are pension promises kept. The discount rate replacement proposal, combined with the other Administration proposals, represent a strong start towards improving and strengthening the defined benefit pension system. We have committed to developing a proposal for fundamental reform and we are working diligently to fulfill that commitment. We look forward to sharing a proposal with Congress in the near future and continuing to work together towards a more secure defined pension system.
1 In years 1 and 2 pension liabilities for minimum funding purposes would be computed using a discount rate that falls within a corridor of between 90 and 105 percent of a 4 year weighted average of the interest rate on a long-term highly-rated corporate bond. In years 3 and 4, pension liabilities would be an average of that calculated using a long-term corporate rate and that using a yield curve. In year 3, the corporate rate would receive a 2/3 weight and the yield curve a 1/3 weight. In year 4 the weights would be switched and in year 5 liabilities would be computed using the yield curve.
2- See Financial Accounting Standard 87.