FROM THE OFFICE OF PUBLIC AFFAIRS

 Chairman McCrery, Ranking Member McNulty, and members of the Committee, I appreciate this opportunity to discuss with you the need to strengthen Americans’ retirement security by measuring accurately pension liabilities.

There is a pension funding problem in America today.  Wall Street firms estimate that current pension underfunding runs to hundreds of billions of dollars.  Pension plan participants are uncertain about their plans’ funding levels and equity markets are unsure of the demands that minimum pension funding may impose on sponsors’ cash flows.  The absence of a clear picture of the extent of defined benefit pension underfunding creates a cloud of uncertainty in equity markets.  Moreover, without an accurate measure of liabilities, the minimum funding rules, which rely upon an accurate measurement of pension liabilities, could lead to insufficient (or excessive) funding of pension promises.

To deal with this challenge, an important step is to develop a more accurate, reliable, and timely measure of pension liabilities. 

 As we go about this task, we must remember that behind all the technical details we will discuss is the retirement security of hardworking Americans.  Our ultimate goal must be to improve pension security for workers and retirees by strengthening the financial health of the voluntary defined benefit pension system that they rely upon.  That system is complex, with many interdependent parts. 

Achieving our objective of secure pensions requires that those pensions be well-funded, that plan sponsors be able and willing to support the defined benefit system, and that the Pension Benefit Guaranty Corporation’s financial integrity be assured.  All three of these groups have an interest in a sustainable program, so all have an interest in getting to the right solution.  In addition, any changes we undertake need to be implemented in a manner that promotes the stability and resiliency of our financial system and financial markets.

 Before proceeding, let me first note that H.R. 1776, the Pension Preservation and Savings Expansion Act of 2003, recognizes the urgency of pension reform and of promoting retirement security.  Its chief sponsors, Congressmen Portman and Cardin, are to be commended for their leadership in this complex, but critical area of public policy.  I would also note that H.R. 1000, the Pension Security Act of 2003, introduced by Rep. Boehner advances principles for improving the defined contribution system that the President set forth last year.  My testimony, however, will focus just on the issue of measuring pension liabilities.

 In our view, overall pension reform that will lead to more secure pensions for American workers and retirees requires three steps:  first, develop a more accurate, reliable, and timely measure of pension liabilities; second, fix the pension funding rules; and third, establish transition rules so as to avoid an abrupt change in firms’ funding plans.

The predicate step to making pensions more secure is to develop a more precise measurement of pension liabilities.  My testimony today will focus on this critical step and, in particular, on the issue of replacing the 30-year Treasury rate as the discount rate used in measuring pension liabilities.  As I will explain, it is critical that Congress develop an appropriate, permanent replacement for the 30-year Treasury rate in measuring pension liabilities.  However, there are many critical questions that need to be answered before settling upon that replacement.  Thus, to give firms the certainty they need to plan for their short-term pension funding obligations, we recommend extending the current temporary corridor for two more years.  At the same time, we need to begin work immediately on getting to that permanent replacement and to dealing with other problems with the current system.

Discounting Future Pension Benefit Payments to Today’s Dollars

Making pensions more secure requires a more precise measurement of pension liabilities.  The amount of pension liabilities determines a plan sponsor’s annual funding obligation.   Without a reliable measure of pension liabilities, plan sponsors may not contribute sufficient funds to their pension plan – or may contribute more than they need to for the obligations undertaken. 

In addition, without accurate, reliable measures neither plan beneficiaries, investors, nor the Pension Benefit Guaranty Corporation know how big the pension obligation may be for a given firm.  Investors that do not have a clear picture of a company’s pension liabilities factor that uncertainty into their credit evaluation of the firm, raising its borrowing costs and lowering its stock price.

In order to get to a more accurate measure of pension liabilities, we need to agree on how to discount future benefit payments to today’s dollars.  After describing why this is so, I will then describe why we believe that we should be working towards a permanent replacement for the 30-year Treasury rate in measuring pension liabilities.

Using a Discount Rate to Measure Pension Liabilities

 Pension liabilities are measured as the discounted present value of the future benefit payments to be made to a pension plan’s participants.  These future benefit payments depend upon numerous factors, including the terms of the particular plan and actuarial and mortality assumptions about plan participants. 

 To get the present value – that is, the cost in today’s dollars of these future payments – these future payments must be “discounted” by some interest rate to show how many dollars today are equivalent to those payments in the future.  As the interest rate that is used to discount future benefit payments declines, the value of those liabilities increases.

A simple example explains this concept.  Suppose someone was offered the choice between $100 today and $110 a year from now.  If that person could invest $100 today at a 10 percent annual return, the two offers would have the same economic value.  If however, interest rates were lower and the person could only earn 5 percent annually, the offers would not have the same economic value.  Instead, the person would need to be offered $104.76 today for the offers to be economically equivalent.  Thus as interest rates decline, the amount of money a pension plan needs today (to have in discounted present value terms the amount of money needed to make future benefit payments) increases.

Background on the Use of the 30-Year Treasury Rate

Federal law sets minimum funding rules for private pension plans.  These rules reflect the complex actuarial work needed to determine the amount of assets that a plan should hold to meet its benefit obligations many years into the future.  One of the most important of these rules is the interest rate for discounting pension liabilities.  Since 1987, the law has used the yield on 30-year Treasury bonds as the basis for this interest rate.  The measurement of a pension plan’s liabilities calculated using this rate is the basis for the federal “backstop” funding rules applied to underfunded pension plans.

Congress chose the 30-year Treasury rate as an approximation of interest rates used in the group annuity market.  In other words, Congress wanted a discount rate that would reflect how much an insurance company would charge a pension plan to assume responsibility for the plan’s benefit obligations. 

Although additional refinements have occurred since 1987, the rate on the 30-year Treasury bond continues to play a prominent role in determining pension liabilities for funding purposes.  Until recently, pension plans could determine the value of their pension liabilities using any rate between 90 percent and 105 percent of the four-year moving average of the yield on 30-year Treasury bonds.  As I will explain shortly, last year, Congress temporarily increased the upper end of this corridor to 120 percent.  Note that the upper end of the corridor produces a larger discount rate and hence a smaller measured liability and a smaller funding requirement.  The lower end of the corridor produces the reverse – a larger measured liability and hence greater required funding.

  However, the Treasury Department does not believe that using the 30-year Treasury bond rate produces an accurate measurement of pension liabilities. 

Why We Need to Replace the 30-Year Rate in Measuring Pension Liabilities

The discontinuation of the issuance of the 30-Year bond – which was part of much needed changes in Treasury financing of government debt – makes replacement of the 30-year rate in pension law necessary.  However, we believe that regardless of whether the discontinuation had occurred or not, there was already growing evidence and concern that the 30-year Treasury was becoming less relevant as a benchmark for use in pension calculations. 

One reason the 30-year Treasury has become less relevant is because of changes in pensions themselves.  The lengthy time structure of the 30-year bond makes it less and less relevant when compared to the shortening time structure of the payments of many defined benefit pension plans.  This shortened time structure is the consequence of the increasing average age of active and terminated deferred participants and the increased proportion of participants represented by retirees.  Using a long-term rate to discount all pension obligations understates the true cost of obligations that will be paid sooner whenever the yield curve is upward sloping (as is true now and is generally the case).

In addition, changes in the Treasury bond market and in financial markets more broadly have made the 30-year Treasury rate less reflective of the cost of group annuities and less accurate as a benchmark for pension liabilities.  The difference between the Treasury yield curve and a high-grade corporate bond curve is not fixed, and that spread is wider today than it was in 1987. 

In response to these concerns, last year Congress provided for a temporary expansion of the upper range of the allowable corridor surrounding the 30-year Treasury for calculating the interest rate used to determine current liability.  This temporary change expires at the end of this year.
 In the absence of a permanent replacement or an extension of last year’s expansion of the upper range, the law will “snap back” to 105 percent as the upper end of the corridor. 

Such an outcome would, in our view, increase the discrepancy between the discount rate mandated in the law and that used to price group annuities.  And since minimum funding rules are based upon measured (current) liabilities, a discount rate that further distorts that measurement will also distort the funding requirements. 

Consequently, we believe that Congress should take action this year to avoid this “snap back.”  And, since firms need to make plans now for the funding contributions they will make next year, we believe that Congress needs to act quickly on this matter.

Finding a Permanent Replacement for the 30-year Treasury Rate

We need to get to a permanent replacement for the 30-year Treasury rate in computing pension liabilities. 

H.R. 1776 offers a permanent replacement for the 30-year Treasury with a measure based upon long-term high-quality corporate bond rates.  We believe that moving from a Treasury full faith and credit discount rate to one based on rates on high-quality long-term corporate bonds could improve the accuracy of measuring pension liabilities.  Pension benefit promises made by private sponsors are not without risk since pension sponsors can and do go out of business.  We think that this risk should be reflected in the computation of pension liabilities.  We also understand that high-grade corporate bond rates are used in group annuity pricing. 

Before Congress selects any permanent replacement for the 30-year Treasury rate, however, it will be necessary to consider several key issues, including the following.

First, different pension plans have different benefit payment schedules, some with quite immediate payment requirements and others whose expected payments are distant in the future.  We know that the yields available on financial instruments are different for these different maturities; typically yields relevant to closer maturities are lower.  Thus the question arises whether an accurate present value measurement of these different benefit payments – some made in the near-term and some in the distant future – should be discounted at rates appropriate to their respective timing.

Both economic theory and current practice in fixed-income markets suggest that the most accurate way to measure the present value of a stream of future cash flows is to match the cash flows occurring at a particular time with a discount rate that reflects the interest rate on a portfolio of financial instruments with the same maturity date.  In this way, the discount rates used would be reflecting the time structure of the cash flows.

 In principle, an accurate measurement of pension liabilities, which is the present value of a series of benefit payments to be made over time, could be more accurately measured if the discount rate used was related to the time structure of those benefit payments. 

Thus, we suggest it would be important to consider whether and how to reflect the time structure of a pension plan’s future benefit payments in determining the appropriate discount rate to use.  At the same time, we recognize that reflecting the time structure of future benefit payments could introduce some added complexity, which would also need to be considered.

Second, under current law, the measurement of both assets and liabilities involves “smoothing” techniques, as do the funding requirements.  Properly measured, pension liabilities are the cost in today’s prices of meeting a pension plan’s future obligations.  If a pension plan’s obligations were to be settled today in the group annuity market, their value would be determined using today’s interest rates rather than an average of rates over the past several years, which is the current practice in measuring current liabilities for minimum funding purposes. 

Using current, unsmoothed interest rates would promote transparency.  An accountant or analyst evaluating a pension plan can readily determine the funded status of the plan if asset values are expressed at current market prices and liabilities are computed using current unsmoothed discount rates.  When either or both of these measures are smoothed, however, it is very difficult to determine the plan’s funded status with any degree of certainty.   While there may be sound reasons to measure current interest rates for discounting purposes using something other than the spot rates on a particular trading day, the current practice of using a four-year average of interest rates raises important questions as to the accuracy of the resulting liability measurement.

Thus, we suggest that consideration be given to whether continuing this practice advances the ultimate objective.  It may be that there are compelling arguments to allow for some smoothing with respect to the funding contributions that plan sponsors make to their pension plans.  We need to carefully review whether four-year smoothing of the discount rate used for purposes of measuring a pension plan’s liabilities continues to make sense.  We also need to consider how eliminating this smoothing could affect the variability of liability measurement, recognizing that under current law the existing use of smoothing still produces volatility in funding requirements.

Third, under current law, pension liabilities are calculated using one discount rate but lump sum payments made by pension plans are calculated using a different discount rate.  The pension liability measurement we are discussing is the basis for funding contributions to be made by plan sponsors – some of which will ultimately fund workers’ annuity pension payments but some of which will be paid to workers in the form of a lump sum.

 Thus, we suggest that it would be worth considering whether and how a permanent replacement for the 30-year Treasury rate in measuring pension liabilities should relate to any possible changes in the discount rate used to calculate lump sum payments.

To this point, my remarks have focused on issues to consider in selecting a permanent replacement discount rate for measuring pension liabilities.  While these issues are critical to the goal of achieving an accurate measurement of those liabilities, there are additional issues unrelated to the discount rate replacement that should also be considered.

Thus, we suggest that, in the process of working towards a more accurate measurement of pension liabilities, the mortality table and the retirement assumptions that underlie the computation of current liability also be evaluated.  There is also the question of whether a sponsor in computing current liability should be allowed to recognize that some retirees opt for lump sums rather than annuities at retirement.  Under current law current liability assumes that all retirees take their retirement in the form of an annuity.  These questions require further study.

I believe that we all need to consider the issues that I have just described to ensure that any permanent replacement to the 30-year Treasury rate results in an accurate measurement of pension liabilities.  The consequence of failing to replace the 30-year Treasury rate with an appropriate discount rate methodology will lead to inaccurate measurement of pension liabilities.  Such an outcome, in turn, will lead to under- or over-funding of pension plans.  The former outcome would make pensions less secure for workers and retirees.  The latter outcome could place an undue burden on plan sponsors by shifting more corporate funds to the pension plan than are necessary to fund the company’s pension obligations.

Interim Steps

 Companies have told us that they need to know what their cash requirements are for funding next year’s funding obligation by the end of the second quarter of this year, but further work is needed to define an accurate measurement of pension liability.  While we have considered alternatives to the discount rate methodology proposed in H.R. 1776, we are not yet to the point of offering a specific replacement.  Yet we agree with those who say that quick congressional action on modifying current law is essential, both because in the absence of such action the law reverts to a discount rate methodology that would be even more distorting than the current rate and because plan sponsors need certainty soon in order to plan for next year’s funding requirements.

To that end, we recommend that Congress enact legislation before the end of this June to extend the short-term interest rate corridor relief that Congress provided in 2002.   We would propose that, for plan years beginning in 2004 and 2005, the upper bound of the interest rate corridor for the deficit reduction contribution continue to be 120 percent of the 4-year weighted average of the yield on 30-year Treasury securities. 

During the time offered by the two year extension, we would look forward to working with Congress and pension stakeholders to work through the complex but critical issues I have described that must be addressed to ensure accurate pension liability measurement and, more importantly, advance our ultimate objective of making pensions more secure.

The change in the method of determining pension liabilities may result in changes in the annual contribution amounts, so transition relief will be required.  In addition, these changes should lead us to consider changes in the current funding rules which would increase the security of the pension promises made to America’s workers and their families.  

I would like to stress the need for quick action on this temporary extension of the corridor.  This action is needed to give companies time to budget for next year’s funding obligation.  At the same time, however, we must also move quickly to deal with the complex questions I have outlined in my testimony.  We need to work expeditiously to come up with a permanent solution, not just for how best to measure liabilities but also for the funding rule changes that are needed.  The testimony you are about to receive from PBGC’s Executive Director Steve Kandarian illustrates the urgency of the work before us.  We look forward to working with you to advance this interim solution and to satisfy the long-term need for accuracy in the measurement of pension liabilities.

Conclusion

 Defined benefit pensions are a valuable benefit and the cornerstone of many workers’ retirement security.  Recent financial market trends have exposed underlying weaknesses in the system, weaknesses that must be corrected if that system is to remain viable in the long run.  It will take considerable time and effort to fix the system.  Developing acceptable solutions will also require the cooperation and flexibility of all interested parties. 

 While we must avoid unnecessary delay, the seriousness of current pension problems and the complexity of the defined benefit system suggest that repairing the system will require time for study and for consensus building.  That is why we recommend that Congress, rather than making a permanent replacement for the 30-year Treasury rate this year, extend for an additional two year period the temporary increase of the pension discount rate used to compute current liability. 

During this two year period government, industry, and participants will have adequate time to develop a set of consistent coherent proposals that will insure that pension funding is adequate, that pension demands on firm finances are reasonable and that the financial integrity of the pension insurance system will be maintained for the workers and retirees that are counting on it for their retirement security.