© 2006, Center on Federal Financial Institutions The Center On Federal Financial Institutions (COFFI) is a nonprofit, nonpartisan, non-ideological policy institute focused on federal insurance and lending activities.

Original release date: August 6, 2006

Douglas J. Elliott

202-347-5770

www.coffi.org

Pension Reform: Summary of Final 2006 Bill

Pension reform legislation is expected to be signed shortly by the President. One of the Senate’s final acts before its August recess was to pass, unamended, the 907-page House pension reform bill, H.R. 4. This represents years of complex negotiation and includes many carefully nuanced compromises, sometimes aimed at the very specific desires of particular firms or industries.

At core, the bill attempts to restore the solvency of the Pension Benefit Guaranty Corporation (PBGC) and improve the funding of corporate defined benefit plans, without pushing too many companies into exiting the traditional pension system. This is a very difficult balancing act, since, short of a bailout, it is virtually impossible to improve PBGC’s finances without shifting cost and risk back towards employers. Naturally, this increases the likelihood that they will freeze or terminate the plans that they offer. Therefore, one’s view of the desirability of the final bill will depend heavily on predictions of future corporate behavior and subjective judgements as to the relative importance of PBGC solvency versus keeping firms in the defined benefit system.

Whatever the merits of the bill, it is very unlikely to restore PBGC to solvency. COFFI’s models suggest that a more reasonable hope is that the $23 billion PBGC deficit grows by less than it would otherwise. Our earlier analysis found that the various proposals might reduce a potential $92 billion hole, in today’s dollars, to $40-50 billion. The author now believes this to be overly optimistic and that a better estimate might be $60 billion. Future stock market and interest rate movements could raise or lower this figure, but it would take extremely favorable circumstances to eliminate the deficit altogether.

COFFI has published 22 papers on PBGC and pension reform, available at www.coffi.org. Readers new to this area may wish to refer to "PBGC: A Primer," "PBGC: Fundamental Questions," and "PBGC: Policy Options," which lay out the background. Details of the original Administration proposal and subsequent bills in the House and Senate are contained in "H.R. 2830 versus Administration Proposals," "PBGC: Senate Finance Pension Reform Bill," "Pension Reform: Issues for the Conference Committee", and "Pension Reform: July 2006 Bill (H.R. 4)." Finally, detailed projections of the effects on PBGC’s finances are given in "PBGC Legislation May Not Restore Solvency," which itself refers to three previous papers explaining COFFI’s comprehensive model of PBGC’s annual cash flows.

This paper focuses on those provisions of the 907-page bill that most directly affect the finances of PBGC’s single-employer program, which accounts for the vast bulk of its deficit. This excludes non-pension provisions, sections related to defined contribution plans such as 401(k)’s, changes to multi-employer rules, and the provisions related to "cash balance" plans.

The major changes can be grouped into four categories:

One of the most significant changes is an increase in the effective funding target from 90% of a plan’s liabilities to 100%. (The funding target is the ratio of pension plan assets to liabilities.) All else equal, this is equivalent to roughly a $200 billion increase in required funding in a system with approximately $2 trillion of liabilities.

Current law has a base of funding requirements originally laid down in the original, 1974, version of the Employee Retirement Income Security Act (ERISA). (For simplicity, we will generally refer only to ERISA, but most ERISA provisions have parallel provisions in the Internal Revenue Code.) Layered on top of this are a set of considerably stricter funding requirements known as Deficit Reduction Contribution (DRC) requirements, which were added after PBGC was faced with a series of significant losses. The DRC rules apply to pension plans that are less than 80% funded or are less than 90% funded and were similarly underfunded in 2 of the last 3 years.

The basic funding requirements are loose enough that many observers have described the current rules as having a funding target of 90% of liabilities, for practical purposes, since this is enough to avoid the DRC regime. The pension reform bill replaces the basic ERISA rules and the DRC rules with a new requirement that plans reach 100% funding within seven years of the 2008 effective date of the bill. Companies are required to reduce the initial underfunding by at least one-seventh per year. Similarly, any new underfunding, such as from a fall in asset values or an increase in pension promises, would need to be eliminated within seven years of when it arose.

The 100% target phases in, starting as a 92% target in 2008, rising to 94% in 2009, 96% in 2010, and 100% thereafter.

Current law allows a company that has contributed more than the required minimum to carry over the extra as a "credit balance" to reduce their cash contribution in some future year. This credit balance earns interest as long as it remains unused. The Administration argues strongly that the credit balance rules have added significantly to past claims on PBGC. Most of the largest claims have come from firms which did not need to make any cash contributions for some years prior to their bankruptcies, often due to the use of credit balances.

The Administration proposed that credit balances be abolished. Instead, Congress retained them, but placed significant restrictions on their use. First, it eliminated a double-counting that exists under current law. This arises because a firm can both use the assets related to a credit balance to reduce the calculated underfunding and then can apply the credit balance directly to substitute for cash contributions that would otherwise be required. Since underfunding levels have a significant effect on required minimum contributions, these credit balance dollars end up doing double duty.

Second, Congress agreed to raise or lower credit balances based on an index of actual returns, rather than always assuming that they rise by a specific interest rate. This addresses an issue that arose when the recent financial bubble burst, as credit balances went up by the specified interest rate even when the actual investments went down significantly.

Under the bill, existing credit balances go into a "carryover" account and any new contributions over the required minimums go into a new "prefunding balance" account. We will refer to the two categories together as "credit balances." As noted, credit balances will also go up or down based on investment returns on the beginning of year balances.

Credit balances are also deducted from pension assets for certain other purposes, such as determining whether a firm falls into the "at risk" category, described below.

Certain plans that are considered to be "at risk" of a future pension default will have to use a more conservative method for calculating their liabilities. This would principally consist of assuming that employees eligible to retire in the next decade would retire as early as possible and would choose to take the more expensive option when choosing between a lump sum distribution and a pension for life.

A plan will be considered to be "at risk" if it is less than 80% funded on a normal basis and also less than 70% funded when using the higher liability calculation applicable to "at risk" plans. The 80% test will be phased in, starting at 65% in 2008 and rising to 80% in 2011.

A "PBGC load" will apply for "at risk" plans that were also "at risk" for 2 of the previous 4 years. Plan years prior to the law’s effective date of 2008 will not count for this determination. The load represents expected PBGC expenses of handling the plan termination and is pegged at 4% of the pension liability plus $700 per participant

Any additional liability resulting from the "at risk" test is phased in at 20% a year. The combined effect of the transition periods is that a firm which remained at 79% funded throughout the period would not have the full "at risk" liability added to its funding requirements until 2015.

Finally, there is an exception for employees at automakers and their suppliers who refused an offer made in 2006 to accept a substantial payment or sweetening of their benefits in exchange for retiring early. (This would apply to offers made by Ford, GM, and Delphi which applied to most of their older employees.) Such employees would not be included as potential early retirees for this calculation, largely eliminating the potential extra cost of being considered "at risk."

Airlines which choose to "freeze" their pension plans would be allowed 17 years to fund their initial pension deficits, 10 years longer than offered to non-airline plan sponsors. In addition, such airlines would be allowed to use a discount rate of 8.85%, which is 2-3 percentage points higher than other firms would be using at current rate levels. Each percentage point of higher discount rate would reduce the calculated pension liability by 10-15%, making a major difference in their funding requirements.

Should interest rates rise by less than 2-3 percentage points cumulatively over the 17 years, there could be substantial underfunding remaining at the end of that period, as the liability calculation switches to that used by other companies. These airlines would then have another seven years to eliminate that underfunding. This follows from the law’s general approach that any additional underfunding arising at a pension plan must be amortized over seven years.

In exchange for these advantages, airlines electing to use the provisions must accept a freeze on the level of the maximum PBGC guaranty for their employees, which otherwise increases by wage inflation each year that a plan remains unterminated. In addition, any airline makes this election and subsequently terminates an underfunded pension plan in bankruptcy will be faced with a higher termination premium when, and if, it comes back out of bankruptcy. This termination premium will be $2,500 per participant in the terminated plan for each of three years, double the rate charged other firms.

Northwest and Delta have pushed hard for this set of provisions and would be the most likely to use them. American and Continental, which are unlikely to make this election, have argued that the provisions would give the other two airlines an unfair leg up. Partially in response to this, a second provision allows airlines an additional 3 years, for a total of 10 years, to fund their initial deficits, even if they do not freeze their plans. However, without a freeze, they do not have access to the 8.85% discount rate.

The existence of a PBGC guaranty could provide economic incentives for employers and employees to agree to benefit levels that are not realistically sustainable, but would be picked up by PBGC if the sponsoring firm were to go bankrupt. The Administration, and others, have argued that this has happened in practice and that rules are required to restrict benefit promises that cannot be met without PBGC insurance. The pension reform bill adds considerably to the existing legal restrictions.

Plans that are less than 60% funded, after subtracting all credit balances, must freeze their benefit accruals. This means that employees would not be entitled to a larger pension for working an additional period of time, even if there is otherwise a contractual commitment to provide that accrual. Such plans also may not pay lump sums to retirees and may not trigger shutdown benefits. Plans that are between 60% and 80% funded are allowed to accrue benefits, but may not increase the richness of the formula. They may pay lump sums, but only up to 50% of the normal entitlement, or the level guaranteed by PBGC, if that is less. The remaining benefit would be paid as an annuity, exposed to any future distress termination of the plan.

None of these restrictions apply if a plan is at least 100% funded when credit balances are taken into account. This could be a very significant provision for firms that have credit balances well in excess of the market value of the assets that were purchased with the original excess contributions.

There is a variable premium under current law of $9 per $1000 of underfunding, however most underfunded plans escape this charge due to an exception for firms that are restricted by the "full funding limit." (See "PBGC: Budgeted Premium Hikes Seem Improbable," for more details.) The bill eliminates this exception, subjecting all underfunded plans to the variable premium.

As in current law, the variable premium calculation excludes unvested benefits from the calculation of underfunding for this purpose. On average, about 6% of pension promises have not yet vested at any given time. Otherwise, underfunding will be calculated in the same manner as in the rest of the bill, as a shortfall compared to the target funding level, which starts at 92% of liabilities in 2008 and rises to 100% by 2011.

The bill also eliminates a 5-year sunset provision on the "termination premium" that was newly instituted this year in the Deficit Reduction Act. This is a charge on any firm that terminates an underfunded plan in bankruptcy, resulting in a claim on PBGC. Should such a firm successfully reorganize and emerge from bankruptcy, it will be charged $1,250 annually for each participant in the former plan for three years.

The companies sponsoring pension plans often place a significant value on being able to plan their future pension contributions well in advance. Current law facilitates this by allowing asset and liability values to be calculated based on multi-year averages, which reduce the variation in any one year. However, this predictability increases the chance that contribution levels will not rise quickly enough when financial market conditions change, such as in the aftermath of the recent financial bubble. The Administration argued in its original proposal that virtually all smoothing of asset and liability measurements should be eliminated and that any needed smoothing of contribution levels should be accomplished in other ways. The Administration’s principal proposed method for smoothing contributions was the seven year grace period to amortize any underfunding.

Congress chose to retain the concept of multi-year smoothing of asset and liability values, but reduced the time period to two years. Thus, the discount rates used to calculate liabilities would be smoothed by averaging historical rates over 24 months. Similarly, plan sponsors could choose to average asset values over 24 months. However, the asset value will be limited to no less than 90% and nor more than 110% of current fair market value.

Firms may elect to switch to a liability calculation that uses only the most current rates. Having done so, they may only switch back with permission of the Secretary of Treasury.

The Administration had also argued that the discount rate used to convert future pension payouts into today’s dollars should be different for different maturities. An obligation one year in the future would be discounted at a one-year interest rate while an obligation twenty years in the future would be discounted at a 20-year rate. This is referred to as utilizing a "yield curve," (see "Yield Curve: A Primer," at www.coffi.org)

Congress instead adopted a simplified version of the yield curve by using three rates, rather than one for each year. The first rate would apply for payments due in years 1-5, the second rate would cover years 6-20, and the third rate would cover the remaining period. The yields would be based on averages of the rates on corporate bonds of the relevant maturities in the top three rating categories, such as "AAA", "AA", and "A". The use of the yield curve will be phased-in at 33% in 2008, 66% in 2009, and 100% thereafter.