Firm Seeks Changes to Proposed Hybrid Retirement Plan Regs

This letter is the response of Towers Watson to the request for public comments on the proposed regulations regarding hybrid plans that were published in the Federal Register on October 19, 2010. In some respects, our comments also extend to the final regulations released at the same time.

Towers Watson is a leading global professional services company that helps organizations improve performance through effective people, risk and financial management. Towers Watson offers solutions in the areas of employee benefits, talent management, rewards, and risk and capital management. Towers Watson employs approximately 14,000 associates on a worldwide basis. Our more than 600 Enrolled Actuaries under ERISA provide actuarial and consulting services to more than 1,700 defined benefit plans in the U.S., including over 350 hybrid plans. We have extensive experience assisting organizations with all aspects of the design, implementation and management of hybrid plans. We appreciate this opportunity to comment. The undersigned have prepared our firm's response with input from others in the firm.

We appreciate the efforts of the staff who participated in drafting these proposed regulations. Hybrid pension plans represent an important segment of the defined benefit plan universe that continue to have appeal to many employers and participants. We believe the guidance for such plans should be comprehensive to dispel the regulatory and legal uncertainties that have, unfortunately, hung over these designs for some time. We also believe it is important for this guidance to be promulgated in the spirit in which Congress clarified the general rules and policies around such plans, so as to encourage their continued maintenance and formation by employers who are still seeking to provide a retirement benefit solution for participants including a defined benefit plan. While the proposed regulations represent substantial progress toward that goal, we believe that, once finalized, the regulations should do more and we provide many suggestions below for doing just that.

REQUEST TO APPEAR AT THE JANUARY 26 PUBLIC HEARING

Towers Watson hereby requests to appear at the public hearing on these proposed regulations scheduled for January 26. As requested, our topic outline is shown below. We intend to spend one to two minutes on each topic.

• The Internal Revenue Service (IRS) and Treasury need to give plan sponsors sufficient time to understand the final regulations when they are eventually promulgated and to determine plan design and administrative changes needed to comply. Due to the extensive nature of these issues, particularly the lead-time needed to design, implement and test changes in administration systems, especially for large complex plans (a category into which many hybrid plans fall), the proposed effective date of plan years beginning on or after January 1, 2012, would barely seem sufficient if the regulations were already final instead of proposed. The history of guidance in this area suggests that regulations seem unlikely to be finalized much before this proposed effective date and, in fact, could require an additional round of proposals. As such, we urge that the effective date be the plan year beginning on or after 12 months following the issuance of final regulations.

• We believe that the maximum permitted 4% minimum interest credit that may be applied to non-equity based rates and the 5% maximum permitted fixed interest credit should each be increased by 100 basis points or more. We will go into further detail below, but our view is that such rates can be supported as more reasonable than those proposed from economic, statutory and practical perspectives.

• The requirements for whipsaw relief were unexpected and are proposed at a time when most plan sponsors have already eliminated these calculations from their plans pursuant to the Pension Protection Act (PPA) Section 1107 required amendment date. This potentially puts plan sponsors in an untenable position and calls for substantial flexibility in terms of compliance alternatives and relief. In addition, these requirements raise many issues that require clarification.

• There are several different designs in place which don't fall neatly into the categories or descriptions contained in the proposal, so it is difficult to determine if and when they are statutory hybrid plans. We request that final regulations provide clear guidance with respect to such plans through examples. We provide some examples and our views on the determinations below.

• The proposal defines a closed group of crediting rates as not exceeding a market return, to the exclusion of all other possible rates. We believe that this approach is needlessly restrictive and inconsistent with the concept of a market rate and PPA. This approach will create substantial compliance costs to modify rates that are not materially different from rates in the closed group and we view this as unnecessary and potentially damaging to the defined benefit system as it will frustrate employers. We urge that the range of acceptable rates be substantially expanded when the regulations are finalized.

• The proposal anticipates § 411(d)(6) relief for certain changes in interest crediting rates and asks for comments on approaches to bring rates that do not meet the standards outlined in the proposal into compliance. We agree that this is a complex issue and urge that substantial flexibility be provided to plan sponsors in this area. We elaborate more below but we believe that the expected value of different rates should be taken into consideration when determining the compliance alternatives available.

Closed Group of Rates

The proposed regulations employ an approach whereby a very limited number of specific rates satisfy the market rate standard, to the exclusion of literally hundreds or thousands of rates that are available in capital markets and therefore by definition are market rates. With respect to equity interest credits, the regulations take a broad and reasonable approach (although we do have some issues with the specifics in this area). The proposed regulations set parameters around the characteristics of Regulated Investment Companies (RICs) that meet the standard and then allow any RIC available in the market to be used. The IRS specifically declined to list specific indices or RICs and appropriately opted to give plan sponsors flexibility to choose from what the market has to offer.

We believe that this approach should be employed for the non-equity interest credits as well. As a first step, we would recommend that the margins on the Treasury-based safe harbors be increased so that they could be viewed as equivalent to the third segment rate, which would be established as essentially the dividing line for what meets the market rate standard. We propose that any bond yield, either corporate or government, that is available in the market and that is expected to have a value less than or equal to the third segment rate be deemed to meet the market rate standard. We believe that such an approach is more consistent with the statute's view of a market rate and would avoid needless changes in existing rates that would numerically meet a reasonable standard of a market rate but would need to be changed merely because they refer to a rate other than one in the closed group.

The IRS could address concerns about employer discretion in determining expected values by requiring a historical demonstration over specific time periods such as 10 to 20 years. The rules might also address concerns about deviations from history by permitting or requiring a revaluation of the rate periodically, such as every five years, with any § 411(d)(6) protection that may otherwise be required being waived for any changes that would be required by such review.

Interest Credit Thresholds -- 3%/4%/5%

We support the issuance of regulations that permit the use of reasonable minimum interest credits, in accordance with the terms of PPA, as well as fixed interest crediting rates. However we believe that the maximum threshold provided for each in the proposed rules -- the 4% minimum and the 5% fixed rate -- should be set at higher levels and we will provide our reasoning below. We support the maximum 3% cumulative minimum with respect to equity-based rates.

Maximum 4% Minimum Rate

General observations

Under IRC Section 411(b)(5)(B)(i)(l), a plan is not treated as having an above-market rate merely because the plan provides for a reasonable minimum guaranteed rate of return or for a rate of return that is equal to the greater of a fixed or variable rate of return.

Although the wording may be interpreted in different ways, certain facts are important to consider. First, the language certainly contemplates the incorporation of a minimum. Second, the statute does not preclude the interest rate after the reflection of the minimum from being above market. The test of whether such a rate is above market appears to be performed before the minimum is reflected, and the minimum is required only to be reasonable. An interest credit that results from a rate not above-market combined with a reasonable minimum is, by statute, defined to be not above-market. Third, the term chosen for the minimum was "reasonable" not "immaterial," "de minimis," or other words that would indicate the minimum was required to have negligible impact.

In light of these observations, we believe that the proposed regulations apply an overly restrictive standard. In the preamble the IRS articulates the thinking behind the proposed 4% threshold, stating "The Treasury Department and the IRS have modeled the historical distribution of rates of interest on long-term investment grade corporate bonds and have determined that those rates have only infrequently been lower than 4% and, when lower, were generally lower by small amounts and for limited durations." Such an approach is fundamentally inconsistent with the concept of a reasonable minimum and instead creates an "immaterial" minimum. In our view this is not what the statute calls for. We believe that a minimum of 5% or higher would better meet the "reasonable" concept in the statute. We offer several points below in support of this view.

We also note that if the IRS is concerned about whether a rate combined with a minimum is above market, then the permissible minimum should vary with the rate in question. For example, a higher minimum should be permissible with a Treasury-based rate than with the third segment rate since the expected value of the Treasury rate is lower and thus can be combined with a higher minimum while not exceeding a market rate. Similarly, if a rate is subject to a maximum then it should also be able to use a higher minimum as the expected value is decreased by the presence of a maximum and this decrease should be able to be restored through a higher minimum without the combined rate exceeding a market rate. Ultimately, this reiterates the point made above that any rate with an expected value that is less than or equal to the third segment rate should meet the market rate standard. There should be substantial flexibility in how this result is achieved in terms of using minimums, maximums and rates outside the closed group.

Historical analysis

The proposed regulations note that historical analysis was performed to determine a minimum rate of interest that would apply only infrequently, and when it did apply would generally have small impact and for limited duration. Although we believe that such a test is inconsistent with the statute, as we have stated above, we performed a similar analysis. We observed the following:

• Since its initial publication, the third segment rate has averaged about 6.5%, rarely dropped below 6% and never dropped below 5%. The pattern of current liability rates was similar, although the rates were somewhat lower, as would be expected by their definition versus the third segment rate. The application of a 5% minimum to the third segment rate would have had no effect, and the application of a 6% minimum would have had the type of minimal effect sought by the IRS. Note that although the period of this analysis is relatively short, it is characterized by historically low interest rates. The effect during a period of normal or high interest rates would be even smaller.

• We studied the average of annual long-term Aaa- and Baa-rated Corporate rates as a proxy for the third segment rate. This series had not been as low as 4% since the 1950s, and had not been as low as 5% since the 1960s. For January through October 2010, this average is 5.49% -- the only year since the 1960s that it has been less than 5.50%.

• We calculated the current balance of a hypothetical account, established with $100 in 1960, that had been credited interest for fifty years at this corporate rate. We applied varying levels of minimum rates to determine the impact that they would have had.

Market-consistent valuation

We used risk-neutral valuation, an accepted technique of financial economics, to value the subsidy that would have resulted from incorporating minimum interest rates in an interest credit that was based on long-term high-quality corporate bond yields.

Since an interest credit based on such corporate rates alone is deemed to be at market, the current value of an account with a notional balance of $100 is $100. Introducing minimum interest rates at higher levels could increase this value. The amount of the increment reflects the probability and magnitude of the effect of the minimum.

We applied these minimum rates on an annual basis. Their value is larger if the account is to earn interest for longer, so we performed the analysis over several different time horizons:

               No minimum   4% minimum   5% minimum   5.5% minimum

This analysis relies heavily on the underlying assumptions, of course, and we sought to select assumptions that would represent long-term representative levels. Key inputs included the following:

• A level of interest rates consistent with the intermediate projection of the 2010 OASDI Trustees report

• Interest rate volatility as implied by swaptions as of a selected representative date

• A credit spread and shape of the yield curve consistent with long-term historical averages

5% Fixed Rate

General observations

We recommend increasing the level of the fixed rate to 6% or higher consistent with our proposed increases to the minimum rate. Support for this recommendation follows.

Historical analysis

We believe that the history of the third segment rate and current liability rates before that and broader data on corporate bonds yields far exceeding 6% for decades before that suggest that a fixed rate of 6% should not exceed a market rate. In our view, any analysis of relevant data will result in the conclusion that a 6% rate is not expected to be above long-term corporate yields for anything other than short periods of time.

We understand that there could be a concern that capital markets could change dramatically and for an extended period, and that it is not inconceivable that 6% could be above corporate yields, although we suspect few would have said this was possible a mere five years ago. We suggest that this concern should not be addressed by promulgating a permissible fixed interest rate that is unduly low. Instead we suggest that the IRS consider revisiting the issue periodically, for example every five or 10 years, and make changes in the regulations to the extent there have been substantial changes in the capital markets. Without such a process, the fixed rate runs the chance of being too high or too low and neither result would be consistent with the intent of the statute. Of course, any required reduction in the permissible rate would have to be accompanied by § 411(d)(6) relief for those plans that exceed the new level.

We also note that the Social Security Trustees, in their annual report, make assumptions regarding the level of future interest rates. Under the intermediate assumptions, the nominal interest rate is projected to rise to 6.1% in 2014 before declining to the ultimate assumed level of 5.7% for 2019. This is the rate for some special-purpose bonds issued to the trust fund, which is set equal to the average yield on Treasury securities of at least four years to maturity. As the third segment rate is a corporate rate for periods in excess of 20 years, the implied assumed level of the third segment rate under these intermediate assumptions would seem to be at least 7% and probably more. This would easily support a view of a 6% or higher fixed rate as not being above market.

Market-consistent values

We repeated the analysis performed above to determine the impact of minimum rates on corporate yields only this time we determined the value of an investment yielding a fixed credit for varying periods of time. The results are shown in the table below, with a value of $100 essentially indicating a market rate. Keep in mind that although this market-consistent valuation assumes that rates will be distributed around the intermediate projection of the Trustees' Report, it does reflect volatility of yields around those points.

                5%           5.5%           6%            6.5%

Disruption to the Defined Benefit System

We encourage the IRS to consider the effect of the regulations on the participants and sponsors of hybrid plans. An unnecessarily restrictive interpretation of PPA would present a number of significant adverse systematic impacts. It is difficult to quantify these precisely.

• The benefits paid to plan participants will decline, of course, if future interest credits are lower. While plan sponsors may compensate to some extent by changing principal credits, reducing interest credits and increasing principal credits reduces one of the key advantages of hybrid plans -- adequate benefit accruals for mobile workers.

• Many sponsors will be required to change plan provisions and administrative systems. A Towers Watson database with over 400 hybrid plans indicates that over 50% of plans with minimum interest credits and almost 30% of plans with fixed interest credits would need to change their interest credits in order to comply with the proposed regulation. Most will need to change (and likely lower) principal credits as well in order to maintain compliance with the accrual rules in light of lower future interest credits. This would introduce a significant burden for these plan sponsors.

• If plan participants do not feel that account balances will be credited with a market rate of interest, they will be more likely to withdraw their account balances on termination which could lead to leakage of retirement assets. Historical data confirm that this problem will exist regardless of what the regulations are, however it will be exacerbated if the regulations require lower rates.

• Sponsors of hybrid plans may be frustrated by another round of required changes and could exit the system in increasing numbers. We note that hybrid plans remain a source of growth for defined benefit plans. As noted in the PBGC Pension Insurance Data Book, over 10% of plans are hybrid plans covering over 30% of total plan participants. Both of these figures have increased substantially in recent years and a reversal of this trend would not, in our view, be a positive development for the U.S. pension system as a whole.

Transition to Market Rate of Interest for Existing Accruals

§ 411(d)(6) Standard

The preamble to the proposed regulations indicates that once they are finalized it is expected that relief from the requirements of § 411(d)(6) will be granted for a plan amendment that reduces a future crediting rate with respect to benefits that have already accrued from a rate that is in excess of a market rate of return under the regulations, but only to the extent necessary to constitute a permissible rate under those regulations. We believe this "to the extent necessary" standard is unduly restrictive and penalizes plans that have waited for comprehensive and final regulatory guidance to be issued before deciding whether and how to make changes to the rate of return currently being credited under their plans.

Under section 1107 of the Pension Protection Act, a reduction in a plan's rate of return was permitted without being treated as a failure to meet the requirements of § 411(d)(6) provided the amendment was made "pursuant to," inter alia, any regulation issued by the Secretary of the Treasury. As another condition for application of the relief was that the amendment had to be adopted no later than the last day of the first plan year beginning on or after January 1, 2009, clearly this relief itself is no longer available for plans that want to react to the final regulations on the market rate standard.

However, we suggest the Treasury consider providing similar relief through a modification of its regulations under § 411(d)(6) that would permit conforming amendments to be adopted no later than whatever deadline is set forth under those regulations more generally for plan amendments to bring plans into compliance with the new standards. That is, we suggest the "to the extent necessary" standard should be replaced with a "pursuant to" standard.

The condition permitting an amendment of otherwise protected benefits only to the extent necessary to constitute timely compliance with a change in law affecting plan qualification is found only in Treasury regulations; i.e., in Reg. 1.411(d)-4, Q&A-2(b)(2)(i). Nothing in those regulations or in the statute to which they relate suggests that the Treasury is barred from crafting other exceptions to the normal restrictions of § 411(d)(6). In this regard, the text following § 411(d)(6)(B) authorizes the Treasury to issue regulations exempting amendments that eliminate or reduce benefits or subsidies that create significant burdens or complexities for plan participants, subject to the condition that such amendment not adversely affect the rights of any participant in more than a de minimis manner. The legislative history to this provision makes clear that it is intended to apply only to an amendment eliminating or reducing an early retirement benefit, retirement-type subsidy or optional form of benefit. A reduction in the interest crediting rate under a hybrid plan does not involve any of these elements of a plan, so the conditions of this special exception should not apply.

Rather, we believe that PPA section 1107 is evidence of Congressional intent to provide flexibility for sponsors to amend their plans to conform to regulatory guidance about the PPA in a manner that eliminates the need to demonstrate that the change is being done only to the extent necessary to conform to those regulations. Given the vagaries of the rulemaking process, it makes little sense to us that Congress would have wanted the relief to only apply to those issues addressed by regulations finalized a relatively short period of time after the enactment of the PPA (i.e., by 2009) and not to regulations finalized at a later time. Rather, we suspect this chronological deadline was enacted with the expectation that all of the regulations needed to implement the PPA would already have been issued by that deadline so as to ensure that sponsors would make the necessary changes by this deadline.

Since we do not interpret the statute or the regulations to necessarily constrict the ability of the Treasury to use its rulemaking authority to make exceptions from the restrictions of § 411(d)(6), we believe it would be appropriate for the Treasury to provide for the "pursuant to" standard for amendments adopted a reasonable time after it promulgates its final, comprehensive market rate regulations. This will not penalize sponsors who reasonably have waited and are still waiting more than four years after the enactment of the PPA for comprehensive guidance about this and other hybrid plan requirements.

Definition of "to the Extent Necessary"

In the event that the "to the extent necessary" standard is retained, the issue of how to bring an interest crediting rate into compliance is an extremely complex one due to the wide variation in practice as well as the options permitted even under the closed group approach of the proposed regulations. The position that § 411(d)(6) protection will be provided for reductions only to the extent necessary to comply forces some difficult questions that need clear answers. And the principles underlying these answers should be applied consistently through the market return proposal and not just for transition. The most fundamental question is what role expected values will play in the determination of the extent of reduction that is necessary. Will the extent necessary be determined as the smallest reduction in expected values of the interest crediting rate? Will it be determined as the smallest possible reduction in the crediting rate? More specifically, will it be determined as the smallest reduction without changing the character (i.e., Treasury-based, corporate bond-based, inflation-based, equity-based or fixed minimum) of the interest credit? Due to the complexity of many interest crediting rates used, it would seem impossible to apply either of the latter approaches in many cases, so that some role for expected value must be established, and should be established in the interest of minimizing the change of both the promise to participants and the costs to employers while still meeting the market rate standard.

This conclusion leads to another fundamental question, which is what is the hierarchy of expected values for the permissible interest crediting rates? The final regulations permit, with some minor limitations, any of the non-equity based rates to be changed to the third segment rate. This implies that the third segment rate is viewed as higher than any of these rates (including permissible minimums). Historical analysis and common sense support this position and we agree with it. However, we don't see the rationale for maintaining rates with different expected values as the ceilings for market rates of return for different indices, and note that the transition issues are made more complex by this disparity. In addition, the potential for cutbacks in benefits is greatly increased. The Treasury-based rates were promulgated in Notice 96-8 at a time when they were the 417 applicable interest rates. Margins were provided on various shorter-term rates to essentially view them as equal to the 30-year Treasury rate. Establishing deemed equivalence for the various acceptable rates was an approach that made a lot of sense then and is one that we recommend be followed now. Now that the third segment rate is essentially the 417 rate, we recommend that the margins on the Treasury-based rates be increased so that they may be deemed equal to the third segment rate. Such a change would simplify transition issues substantially and would also be advantageous to participants because it would enable the continuation of rates such as "30-year Treasuries + 1%" which would no longer meet the market rate standard even though such a rate would not be expected to exceed the third segment rate.

Neither the final regulations nor the proposal contain any provisions for changing between equity rates and non-equity rates. We believe that equity rates in general would have a higher absolute expected value than the other rates but not necessarily a higher risk adjusted value. Equity rates come with substantially more volatility as well as the possibility of negative credits for certain years. In light of these factors, we understand if there is a reluctance to grant § 411(d)(6) relief for moving to or from an equity rate. However, we recommend that § 411(d)(6) relief be provided for changing from an equity-based rate to the third segment rate in recognition of the value of receiving a high-quality, long-term corporate bond yield that is reset each year without risk to principal. In no instance should there be a requirement to move to an equity rate from a non-equity rate in order to obtain § 411 (d)(6) protection as requiring that participants be subject to the volatility of an equity rate does not seem justified.

Before getting into our specific recommendations for a transition approach, we note that the decision to use a closed group of permissible interest crediting rates greatly adds to the complexity as well as the number of transitions needed. We reiterate our request that this approach be reconsidered and that a much broader spectrum of rates be permitted to meet the standard of not exceeding a market rate.

That being said, our recommendation for a transition approach with § 411(d)(6) protection appears below. We recognize that this approach is very detailed and we would normally not suggest an ongoing process that would be this extensive. However, we believe in this instance it is critical to articulate a process that addresses the many varieties of interest crediting rates that are in existence in a reasonable manner that addresses the concerns of plan participants while providing some flexibility to plan sponsors. We note again that this process could be greatly simplified if our suggestions above are adopted.

Finally, the regulations should provide that plan sponsors will be deemed to have complied with the market rate standard under IRC § 411(b)(5) for the period from its statutory effective date through the date immediately before the effective date of the final regulations based on a reasonable interpretation of the statute. This is similar to the standard adopted by the IRS in Notice 2010-77 in which it announced that its review of plans submitted for a determination letter after January 31, 2011 with respect to matters regarding IRC § 411(a)(13) and 411(b)(5) addressed in proposed regulations would be based on a reasonable interpretation of the statute. We suggest that the purposes for which this standard applies should be broadened through the regulations to apply for all purposes in which an interpretation of these statutory provisions is relevant. Use of a reasonable interpretation of the statute for these purposes seems to property balance the interests of plan sponsors and participants during the period before comprehensive regulatory guidance is applicable.

Participant Notice of a Required Reduction in the Rate of Return

We believe that the final regulations or related guidance should relieve sponsors of the need to furnish a 204(h) notice to participants in connection with an amendment to reduce the interest crediting rate under a hybrid plan if, without regard to that amendment, the plan no longer provides any ongoing principal credits or other benefit accruals after the effective date of the amendment. An example of a situation in which this would apply is a frozen cash balance plan that is being amended to reduce its interest rate to a market rate as set forth in final regulations.

Reg. 54.4980F-1, Q&A-6 defines an amendment to a defined benefit plan that reduces the rate of future benefit accrual as one that is reasonably expected to reduce the amount of the future annual benefit commencing at normal retirement age for benefits accruing for a year. We believe the phrase "for benefits accruing for a year" is meant to generally limit the need for a 204(h) notice to a reduction that affects benefits accruing at a time after the effective date (or later adoption date) of the amendment that will potentially trigger the need for such a notice. If it had been intended to mean benefits accruing in any year, including those accruing prior to the effective date of the amendment, it seems that the phrase would not have been necessary since the reference to the amount of future annual benefit commencing at normal retirement age would have already imparted that meaning to the standard in Q&A-6. Also, this interpretation seems most consistent with a plain reading of the statute itself -- which focuses on reductions in the rate of future accruals.

Reg. 1.411(b)(5)-1(e)(3)(i) makes the point that the right to future interest credits not conditioned on future service generally is a protected benefit as defined in Reg. 1.411(d)-3(g)(14). The list of protected benefits in that regulation is short. As the right to future interest is not an early retirement benefit, a retirement-type subsidy or an optional form, it must have this status as a part of the benefit that has already accrued.

This interpretation of the applicable regulations supports the position that a 204(h) notice would not be necessary for a frozen plan for the reduction in the rate of return described above as the right to the future rate of return was accrued prior to the effective date or adoption date of the amendment. The Treasury or the IRS should clarify this in applicable guidance.

Equity-Based Rates

We applaud the service for providing the flexibility to provide interest crediting rates that are based on equity returns. The proposed regulations raise the question of how section § 411(d)(6) applies to a crediting rate equal to the return on a RIC if that RIC ceases to exist. This is one of several questions that we believe need to be answered in order to implement equity-based returns for new plans and to bring those for existing plans into compliance.

We believe that the issue of changing equity rates should be addressed more broadly and not just with respect to situations where an RIC ceases to exist. In particular, our view is that plan sponsors should generally be given § 411(d)(6) relief for changing equity-based rates in a variety of situations. The situations in which such relief would seem appropriate are those that might typically cause a plan sponsor to consider a change in the investment of assets in one of its qualified plans, including, but not limited to:

• The RIC ceasing to exist

• An extended period of underperformance by the RIC relative to targeted benchmarks

• A change in management or investment philosophy by the RIC

• Substantial departure from stated policies by the RIC

• Plan mergers

We understand that this involves a delicate balance of sponsor and participant interests but we encourage the IRS to provide employers with flexibility in this area so that participants can benefit. As discussed above in the transition section, the expected return would be a range and would need to be supported by historical data and capital market simulations. The IRS might publish a list of valid reasons and establish a process for approval of changes for reasons outside of that list. For other situations, existing accrued benefits would need to be protected.

For plans that do use the actual return on plan assets, we recommend that the regulations explicitly provide that changes to a plan's asset allocation, investment managers or other aspects of the investment strategy that are made in the normal course of the financial management of the plan are not viewed as changes in the plan's interest crediting rate. Furthermore, the regulations should state that using the plan's actual asset return as the interest credit is not viewed as making plan benefits subject to employer discretion.

Lastly, with respect to any equity rate we recommend that the interest credit be permitted to be based on the appropriate return for the prior year, and that the pro rata application of this rate to computation periods less than a year be permitted.

Blended Rates

The final regulations permit the application of different interest credit rates to "different predetermined portions of the accumulated benefit". We request clarification on what is intended by predetermined portions. We believe that the intent is that a predetermined portion could be either a specified percentage of the account value at specified points in time (e.g., 50% of the account value at the start of each stability period receives credit rate 1 and the other 50% gets credit rate 2) or a sub-account stipulated in the plan (e.g., opening account balances or all accounts accrued before a specified date). Please confirm that both of these constructs are permissible.

Assuming that the first construct above is permissible, we request that the regulations clarify that the rates can be combined into a single combined interest credit rate that would meet the market rate standard. It would be beneficial to plan sponsors to have it explicitly stated in the regulations so that there is additional plan design flexibility without the additional plan administration that would be required to actually maintain a separate interest credit on portions of the account. We understand that this additional administration would be required if the different rates had different cumulative minimums or maximums placed on them as this would not be mathematically equal to combining them into a single rate. However, if rates had different annual minimums or maximums then they could be combined into a single rate as long as the limits are applied to each rate before they are combined.

Averaging

We request that the regulations permit interest credits that are the average of otherwise permissible credits over a number of stability periods, not to exceed five years. The use of an average should not impact whether a rate meets the market rate standard as over the long term, the averaging has no impact on the expected level of the interest credit. Averaging has the positive result of producing more predictable rates. This is advantageous to participants in terms of retirement planning as well as to plan sponsors in terms of budgeting costs. We also suggest that averaging of rates be explicitly mentioned as an approach that would not cause an assumption to be viewed as unreasonable for the purpose of annuity conversion requirements. We note that the proposal seemingly permits averaging over the five-month potential lookback period, and the ability to use any version of the segment rates permits the use of a 24-month average rate. At a minimum, we recommend that other rates be permitted to be averaged over the same 24-month period as the segment rates.

Period Used to Determine Interest Credit

In general, the regulations require the rate of interest being used to be the effective rate that applies for each current interest crediting period specified in the plan. For safe harbor interest credits, as an alternative, the rate credited can be fixed for a "stability period" (of up to a year), and can be determined based on a "look-back month" before the stability period begins, using the same rules that govern look-backs and stability periods for purposes of the "applicable interest rate" under IRC § 417(e) for lump sums under non-hybrid plans. Under the alternative approach, it appears that interest cannot be determined on the basis of a period that is shorter than one month.

We recommend that the regulations be modified to permit plans using safe harbor rates to determine the appropriate rate with reference to a period that is shorter than one month without regard to the interest crediting period used by the plan. For example, the plan should be able to use a one-day rate for long-term investment grade corporate bonds even if the interest crediting period is a plan year.

2. Whipsaw Relief

Many plan sponsors have refrained from amending their plans to eliminate whipsaw because they wanted to make needed changes all at one time after all the rules are known, so as to best assess their options. This is an eminently reasonable approach that should be accommodated by extending the period during which whipsaw calculations may be removed without violating the anti-cutback rules. Other plans have been forced through the determination letter process to add "annuity whipsaw" calculations to their plans, even after PPA, and the regulations clearly indicate that no such calculations are necessary. These plans should clearly be given an opportunity to remove such provisions without violating anti-cutback rules.

Early Retirement Subsidies

The proposed regulations contain criteria for whipsaw relief for non-single sum forms of benefit (i.e., benefits other than a single payment equal to the then-current balance of the hypothetical account or the accumulated percentage of the participant's final average compensation). It is not clear what the intention of these requirements is nor what the implications would be of meeting the whipsaw requirements for one form of benefit and not another. Our hope is that the intention of this section is to confirm that annuities prior to normal retirement can be determined directly from the account balance as opposed to requiring them to be based on the accrued normal retirement benefit. Such confirmation would be critical as the vast majority of hybrid plans determine benefits in this manner. However, the requirement that undefined reasonable assumptions be used and that the value of the optional form be equal to that of the account balance could have consequences that are not desirable in our view and, we suspect, were not intended.

While we of course would not object to a requirement to use reasonable assumptions, having such a requirement together with the required actuarial equivalence applying at all ages suggests that there is no ability to subsidize benefits at any age or in any form. In particular, there is no ability to subsidize early retirement annuities relative to the normal retirement benefit or the lump sum. This is not a desirable result and would be a severe restriction on hybrid plans that would not apply to traditional plans. A traditional plan can pay an unreduced early retirement benefit and/or subsidize the Qualified Joint and Survivor Annuity (QJSA). The value of such a subsidy need not be reflected in any lump sum offered by the plan. Under the proposal as written, a hybrid plan would appear to have no ability to do this especially if the reasonable assumption requirement is interpreted as requiring the same set of reasonable assumptions to demonstrate compliance for all forms and ages. Even if this were not the case, it would typically be difficult to find different assumptions within the reasonable range that would both support (1) actuarial equivalence of an account at, say, age 55 to an unreduced normal retirement benefit, and (2) actuarial equivalence of the normal retirement benefit to the account, and also meet the criteria that the account could not be reduced other than for specified reasons.

We respectfully suggest that the proposal be modified to clarify that subsidies are permitted and need not be included in the account balance or lump sum. This can be accomplished by simply modifying the proposal to require that optional forms and immediate annuities be at least the actuarial equivalent of the account using reasonable assumptions as opposed to equal to the actuarial equivalent. Alternatively, this could be remedied by simply stating that it is permissible to determine early retirement benefits and optional forms from the account balance rather than the accrued benefit and rely on existing rules, which apply to all plans, regarding the relationships between such benefits and the accrued benefit to ensure compliance. For example, the early retirement benefit would be limited to the accrued benefit (determined solely for this purpose by using reasonable actuarial assumptions about future interest credits and annuity conversions) regardless of what the account produced as an annuity. If precluding hybrid plans from subsidizing early retirement annuities or optional forms was indeed the intent, we would strenuously object as there is no basis in the statute or any good policy reason to impose such restrictions on hybrid plans.

The approach of relying on existing rules to limit subsidies would be preferable as early retirement subsidies can take many forms in hybrid plans. For example, some plans might convert the account into an annuity at early retirement ages using a higher interest rate than at normal retirement, subject to the limitation that the resulting annuity is never more than an accrued benefit. We believe that this approach should not be problematic; however, the higher interest rate, in isolation, might not be viewed as reasonable. We would disagree with this latter point and would characterize the excess of the interest rate over that used to convert the account at normal retirement as an early retirement subsidy that should be disregarded when determining if the resulting benefit is permissible. Other plans might use a flat factor to convert the account to annuities at all ages. This is similar to providing an unreduced benefit (although not quite as generous as the normal retirement benefit would have additional interest credits in it) and should be permitted, yet it might be difficult to find reasonable assumptions that could support this approach (and impossible to do so if the same reasonable assumptions were required to be used at all ages).

Flat Factors

The use of flat factors (to convert the account to an annuity) in hybrid plans raises another issue that is potentially created by the promulgation of requirements for whipsaw relief after many plans have removed such calculations. Plans use flat factors for a variety of reasons including ease of understanding for plan participants, simplification of plan administration, benefit and cost predictability and provision of early retirement subsidies. We believe that regulations should accommodate the use of flat factors. However, some plans may have difficulty meeting the reasonable assumption requirement and should be provided § 411(d)(6) protection to change their factors to meet this requirement and retain the whipsaw relief that they have already availed themselves of without knowledge of this requirement.

More importantly, for the majority of plans that can meet the reasonable assumption requirement, there is no assurance that the factors will continue to meet this requirement in the future. Changes in capital markets or mortality improvements may call into question the reasonableness of the factors over time. In the event that this occurs, sponsors should have the ability to amend the factors to ones that meet the reasonable standard. Because of § 411(d)(6), benefits accrued to the date of change may need to be protected using the previous factors; however, the continued limited use of these factors that are not reasonable would seem to preclude the sponsor from continued whipsaw relief. Even though the impact of the old factors might wear away quickly, once whipsaw relief is lost it would seem to be lost forever. We believe that this is not the correct approach and that the regulations should provide that maintaining previous factors as may be required by § 411(d)(6) does not cause a plan to lose whipsaw relief that it would otherwise qualify for. We note that this would be similar to the approach taken in the proposed regulation with respect to changes from one acceptable interest crediting rate to another whereby the new rate is not deemed to exceed a market rate solely because the required preservation of the account accrued at the time of change increased with the old rate is applied.

Separate Criteria for Relief

With respect to whipsaw relief being applicable separately to lump sums, early retirement annuities and optional forms, we request guidance on what it means for a plan to meet the requirements in one situation and not another. If our suggestions above are followed and the relief for annuities simply becomes a statement that it is acceptable to base these amounts on the account balance then we believe that this issue no longer exists. If this is not the case, however, does failing to meet the relief criteria for any form of benefit mean that no benefit (regardless of form) can be paid without performing a whipsaw calculation? Would this apply to an individual or on a plan-wide basis, such that failure of any form for any individual would require whipsaw calculations to be performed for all participants? These results seem unreasonable and we would suggest that any requirements to perform whipsaw calculations be applied narrowly to those instances where the criteria are not met.

Reasonable Assumptions

We welcome the design flexibility afforded by the proposal to use reasonable assumptions to meet the criteria for whipsaw relief. Such an approach is clearly preferable to mandating a single set of assumptions to be used. However, we believe that a framework for evaluating the reasonableness of assumptions should be included so that sponsors may design their plans with confidence that they will comply.

We suggest that the framework for reasonability be based not on rates in a particular year but rather on long-term reasonable estimates, supported by historical data, of what the rates will be. As discussed above in relation to the use of flat factors, what will generally be viewed as reasonable can change over time but this change should be a lot more gradual than the year-to-year fluctuations in any particular prescribed interest crediting rate. Other rates that are prescribed in pension regulations should also be viewed as reasonable, such as the rate used to adjust IRC Section 415 limits and the rates used to normalize benefits for nondiscrimination testing. It would seem inappropriate to require the use of an interest rate for certain purposes yet view that rate as unreasonable for other closely related purposes.

Guidance on reasonable assumptions is particularly important for the issue of determining benefits at normal retirement. Since actuarial equivalence between the accrued benefit and the account at normal retirement is the key to whipsaw relief, sponsors will want assurances that the basis on which they perform this calculation will be viewed as reasonable. Clearly the Internal Revenue Code (IRC) Section 417(e) assumptions would be reasonable but the regulations allow for the use of any reasonable assumptions. We request that the IRS provide guidance that will assist sponsors in evaluating what alternative assumptions would also be viewed as reasonable for this purpose. Note that the issue of permanently losing whipsaw relief if assumptions ever become unreasonable, as discussed above relative to flat factors, is present here as well. Again, we would request that the regulations specify that providing any required anti-cutback protection when changing from assumptions that are no longer deemed reasonable to assumptions that are will not cause a plan to lose whipsaw relief.

Plans That Eliminated Whipsaw but Do Not Meet New Criteria

Many plans have eliminated whipsaw calculations due to the PPA section 1107 deadline to do so with anti-cutback protection and the absence of any requirements for that relief other than being a lump sum-based plan. Some such plans may find that they have difficulty meeting the newly proposed requirements for that relief. Such plans should be afforded the ability to amend their plans to qualify for the relief with the provision that any required anti-cutback protection resulting from the change would not disqualify them for whipsaw relief. Furthermore, the final regulations should explicitly state that there are no compliance issues associated with plans eliminating whipsaw for the period prior to the effective date of the final regulations as they were not subject to the conditions in these proposed regulations and were covered with respect to § 411(d)(6) by PPA Section 1107.

"Greater of" and "Sum of"

We do not believe that there should be a distinction between the greater of and sum of approaches as contained in the regulation. Both approaches will produce the same benefit assuming that the same options are offered under each formula so the semantics of how a benefit is described should not impact its determination. In our experience, sum of situations generally arise when there is a greater of formula but the traditional formula is not available as a single sum. In such a situation we would recommend that the greater of approach be followed and that 1) the current account be paid as the single sum under the lump sum-based formula, 2) the single sum value of the traditional formula accrued benefit be determined using 417(e) assumptions, and 3) the remaining traditional plan accrued benefit be determined as 1 minus the ratio of the account to the single sum value of the traditional plan accrued benefit, multiplied by the traditional plan accrued benefit. Early retirement benefits would then be determined under the terms of the traditional plan formula based on this remaining accrued benefit. We also believe that a plan should be permitted to provide that similar calculations be done using the early retirement benefits in order to preserve additional early retirement subsides that were contained in the traditional plan.

In situations where a single sum of the entire benefit is available we see no reason to distinguish between greater of and sum of formulas. In particular, we see no reason to apply 417(e) factors to the account balance to determine an annuity offset.

We also request that the final regulations include a statement that nothing in the regulation would require a single-sum of the entire benefit to be offered in a greater of or sum of situation.

To the extent that a distinction between greater of and sum of formulas is retained, we request that examples be provided to illustrate the intent and application of these provisions as they were unclear to us.

Pension Equity Plan (PEP) Issues

The preamble states that a PEP that does not credit interest after termination and instead converts the then-current account into immediate and deferred annuities is a lump sum-based plan during the period of PEP accruals and is not a lump sum plan thereafter. As a result, the preamble says that such a plan will be eligible for whipsaw relief during the period of PEP accruals but ineligible for such relief afterwards. Please confirm that such a plan is eligible for whipsaw relief if it pays the account balance immediately upon the cessation of PEP accruals (assuming it otherwise meets the criteria for such relief). Furthermore, we would also request clarification that such relief would not be lost if the plan afforded the participant a reasonable period within which to make a lump sum election (such as six months), during which time a lump sum equal to the PEP balance could be paid.

The proposed regulations are clear that a PEP that does not credit interest ceases to be a lump sum-based plan upon cessation of accruals. We request confirmation that future annuities determined under such a plan as the actuarial equivalent, using reasonable assumptions, of the account balance at separation are acceptable -- that is, they are either covered by or do not require any whipsaw relief. We also request confirmation that this situation applies even when the conversion process involves projecting an account to future dates with interest that is locked in at cessation of accruals. Such a plan does not explicitly credit interest nor maintain accounts after cessation of accruals so we believe this approach is no different than using an actuarial factor to convert the account into annuities and thus should not continue to be viewed as a lump sum-based plan. In such a situation we would also request confirmation that the actuarial factors being used in this process are subject to a reasonableness standard but not the market return standard.

One of the proposed requirements for whipsaw relief is ". . . the balance of the hypothetical account or accumulated percentage of the participant's final average compensation may not be reduced except as a result of . . .". Please confirm that the requirement for PEPs is only that the percentage of the participant's final average compensation not be reduced, other than for permitted reasons. Declines in final average compensation can possibly cause declines in the dollar value of a PEP account but this should not be a reason to withhold whipsaw relief.

3. Normal Retirement Value Plans

The proposed regulations ask for comments on whether a defined benefit plan that expresses a participant's accumulated benefit as "a current single-sum dollar amount and that does not provide for interest credits" be excluded from the definition of a statutory hybrid plan. We have strong convictions as to what the characteristics of such a plan must be for the question to be relevant. We believe the description above applies to a PEP that offers a lump sum at termination and does not change this amount even if the lump sum is not taken until normal retirement age. We believe that such a plan is a statutory hybrid.

There are other types of plans that express benefits as a single-sum dollar amount at normal retirement rather than currently. Such plans also do not credit interest because there is no current amount on which to credit interest. We believe it is clear from the final regulations that such plans are not statutory hybrids. We want to make sure that the question does not pertain to these other designs and request that the final regulations confirm that such designs are not statutory hybrid plans. The distinguishing characteristic of such designs, which we will refer to as normal retirement value plans, is the presence of a defined, known value at normal retirement age.

The names of such plans may vary, but we commonly work with two types. We refer to the first type as a stable value plan and such a plan will typically define the annual accrual as a normal retirement value payable at normal retirement as a percentage of the current year pay. It is a career average pay plan. A dollar amount is credited to the normal retirement value each year. Interest credits are not part of the plan because the dollar amounts are defined as payable at normal retirement rather than currently. If amounts are payable prior to retirement, there is an early retirement reduction applied to the value. Applicable and appropriate 417(e) adjustments are made. Participants who leave prior to normal retirement age can receive the known normal retirement age amount by waiting until normal retirement age to commence benefits.

The second type of plan we refer to as a deferred PEP, and such a plan would define a percentage of final average pay to be credited to a normal retirement value. It is a final average pay plan. Again, if payment was taken prior to normal retirement, the value would be reduced by an early retirement factor. Applicable and appropriate 417(e) adjustments are made. Participants who leave prior to normal retirement age can receive the known normal retirement age amount by waiting until NR age to commence benefits.

Lump sums are typically offered under either type of plan, but need not be. Annuities under either plan are typically determined as the actuarial equivalent of the value that would be payable at the age in question. The accrued benefit under either plan is typically the known normal retirement value converted to an actuarially equivalent annuity at normal retirement age.

The current construct of the final regulations as to what constitutes a statutory hybrid is excellent as long as the proper interpretation is made. In our view, normal retirement value plans are clearly not lump sum-based plans nor do they have the same effect as lump sum-based plans. The plan document does not state the value as a current (prior to normal retirement age) accumulated percentage or a current lump sum amount. Nor does it express accruals to the value in these terms, instead defining those as known amounts at normal retirement. The plans function like a traditional pension plan except that benefits are defined in terms or a normal retirement value instead of a normal retirement annuity.

Normal retirement value plans do not require a new safe harbor to escape suspicion of age discrimination; they are exactly the same as their career average and final average pay annuity counterparts in regard to this issue. Accrual rules are passed on known normal retirement age percentages/dollar amounts; again, they are exactly the same in this regard as their career average and final average pay annuity counterparts. Normal retirement value plans do not require whipsaw relief and are subject to 417(e); again, they are exactly the same in this regard as their career average and final average pay annuity counterparts. We respectfully request that the service confirm in the final regulations that normal retirement value plans are not statutory hybrid plans.

4. Conversion Issues

We appreciate the efforts of the IRS to accommodate a "set and forget" approach to determining opening balances for a conversion to a hybrid plan. We also appreciate the interest of the IRS in extending such an approach to forms of benefit other than a lump sum. We believe that limiting this option to the lump sum (and then only in very limited circumstances) will result in very few plan sponsors availing themselves of this alternative. Extending the option to other forms of benefit could make it more attractive to sponsors, but only if this is done in a manner that is practical. We believe that approaches should be permitted -- for lump sums as well as for other optional forms -- in which the expectation is that the opening account will produce benefits that are at least as great as the prior plan benefit, rather than a mathematical guarantee of such a result. While the set and forget approach for lump sums does not explicitly establish such a guarantee, the multiple criteria required for its use essentially does.

If the IRS envisions strict criteria designed to all but guarantee that amounts will not be less than under the prior plan then we suggest that the attractiveness to plan sponsors of any approach that would meet this standard would be so limited as to make it not worth the time and effort associated with designing. Alternatively, if the IRS is willing to move off the lump sum approach and make it more flexible, then we believe that approaches in which there are reasonable relationships based on expected values between assumptions used to create opening balances, interest crediting rates and assumptions used to annuitize balances could be designed that would enable the conversion protections to be satisfied by definition. Such an approach would be beneficial to plan participants in that there would be an improved ability to understand benefits and plan for retirement through consolidation into a single formula. This approach would also be beneficial to plan sponsors as it would reduce the burden of plan administration.

The proposal specifically asked the question about early retirement subsidies in the prior plan benefit. We offer three broad approaches to this issue but have not done a detailed analysis. We would be happy to do so if the IRS indicates the flexibility discussed above. The three approaches we might suggest would be to permit early retirement subsidies to be built into the account balance, permit the value of the subsidy on the prior plan benefit at transition to be added to the account at retirement but characterize it as an early retirement subsidy (so that the amount is ignored for compliance purposes and the amount would decline as the participant continues to work) and permit the prior plan early retirement factors to be applied to the normal retirement benefit produced by the account (determined solely for this purpose by using reasonable actuarial assumptions about future interest credits and annuity conversions). Any of these approaches present issues as the regulations are now proposed, Such issues would need to be addressed and we would be happy to assist in this process if the IRS indicates the intent to draft more flexible proposals that will be attractive to sponsors.

Note that we understand the position that the IRS has taken with regard to this issue. We are merely commenting that if this very strict adherence to the no reduction concept is retained, we do not believe that a "set and forget" approach is practical and worth the effort of attempting to design.

5. Post-Normal Retirement Accruals

The proposed regulations provide that lump sum-based benefit formulas must provide actuarial increases, or properly suspend benefits, if the sum of interest credits and principal credits is not otherwise sufficient to provide "any required actuarial increases" under IRC 411(a)(2). We agree that statutory hybrid plans are subject to the same requirement to pay benefits at Normal Retirement Date (NRD), to properly suspend them, or to provide actuarial increases, as other plans.

Regulatory guidance has not been provided with respect to many aspects of suspension of benefits rules, most importantly including what level of increase would meet the "required actuarial increase," and thus actual practice is mixed, both with respect to lump sum-based benefit formulas and other formulas. For example, interest rates that form the basis for actuarial increases in non lump sum based benefit formulas vary widely. In addition, there is wide variation in plan language relating to suspension of benefits. Some plans provide explicit language relating to suspension of benefits, including references to suspension of benefits notices. Other plans suspend benefits by virtue of the manner in which the benefit commencement date is defined (i.e., not before termination of employment) and the manner in which the benefit then payable is described (i.e., the plan's normal accrued benefit formula, with no mention of actuarial increases), but without using the word "suspension" or discussing suspension of benefits notices (although such plan sponsors distribute suspension of benefits notices if they do not believe that their plans include provisions that meet the requirement to provide a reasonable actuarial increase).

More specific guidance as to what constitutes a "required actuarial increase" would be helpful. However, any such guidance should be prospective only in recognition of the varying practices plan sponsors have adopted in a reasonable good faith effort to comply. In particular, plan sponsors whose plans provide that benefits will not begin before termination of employment, and who did not provide suspension of benefits notices because they reasonably believed that their plan design (e.g., continued interest credits and principal credits post NRD) provided the required actuarial increases, but whose interest crediting rate is judged under final regulations not to provide a reasonable actuarial increase, should be permitted to begin issuing suspension of benefits notices prospectively, without a need to provide additional actuarial increases for periods before the final regulations become effective. In addition, should any plan, as a result of new guidance, be required to retroactively provide actuarial increases not previously provided by the plan (because they were reasonably expected not to be required), the plan should be permitted to provide only the greater of actuarial increases and continued accruals rather than providing both, merely because the plan document did not discuss actuarial increases, and therefore did not provide that continued accruals would be reduced by actuarial increases.

With respect to the level of "required actuarial increases," we recommend the following:

We understand anecdotally that the IRS's position is that, when performing IRC 410(b)/401(a)(4) testing, and testing that the plan meets IRC 411 accrual rules, current values of interest credit and annuity conversion rates must be assumed to continue for all years into the future. We disagree with that position, in that there may be, and have been, economic climates in which there are fairly short-lived, historically aberrant, interest rate environments that are not expected to continue. Since plans with variable interest credits and conversion rates do not actually accrue benefits that come with current rates locked in for all future years, we believe it inappropriate to make such an assumption for testing purposes, particularly when the actuary's best estimate of the long-term value of the index, as reflected in the IRC 430 valuation for the year, is very different. We ask that IRS institute any such requirement only via a proposed regulation exposed for public comment. In addition, the treatment of equity returns would seem to require additional consideration. While the ability to apply a minimum of zero as the assumed future equity return in years in which it is negative is helpful, it still requires the use of an unreasonable assumption about the future (0% equity returns), and imposes no rational limit on the assumption in a year when equities advance rapidly.

We also request that IRS provide guidance on the application, if any, of this principle to IRC 415 limits. For example, if benefits begin before age 62, the IRC 415 limit must be actuarially reduced using 5% interest and 417(e) mortality, or the plan's actuarial equivalent reduction, whichever produces the smaller 415 limit. In a cash balance plan, the plan's reduction for early commencement depends on the interest crediting rate. If a participant retired when the plan was crediting 20% equity returns, if the plan were to be assumed to credit 20% in all future years, the 415 limit would be drastically (and inappropriately) reduced.

Note that we do not believe that the most appropriate solution is to put a cap on assumed equity returns for accrual rule testing, discrimination testing, and IRC 415 "plan actuarial equivalence" purposes, but rather to acknowledge that longer-term best estimate assumptions should be used for variable rates for all purposes, not just funding purposes. Alternatively, an approach similar to that used for setting assumed future rates upon plan termination should be used (e.g., assumed rates equal actual average rates over preceding five years).

7. Participant-Directed Accounts

We believe that the regulations should be supportive of statutory hybrid plans in which the participants choose the interest credit that they will receive from a variety of options. We believe that life cycle funds would seem a reasonable option to offer but do not believe such an option (or any specific option) should be required. We do not believe that any § 411(d)(6) issues should exist as a result of the participants making periodic elections, as the elections and associated changes in interest credit rate are not plan amendments and thus are not covered by § 411(d)(6). Furthermore, the regulations should explicitly state that no forfeiture occurs as the result of participants' elections to change their interest crediting rates.

We note that issues will arise when an index or RIC becomes unavailable, as well as when the plan sponsor wants to change the menu of options available. These issues are virtually identical to those discussed above pertaining to equity interest credits and we would propose the same approaches. In general, we would propose that § 411(d)(6) relief be provided when an option is eliminated for specified reasons that are generally in the best interests of plan participants. We also believe that § 411(d)(6) relief should be afforded in situations where the option that is eliminated is replaced with a new option that has a similar asset allocation and has an expected rate of return that is no less than the option being replaced.

As a matter of plan design, we would expect that plan sponsors that employed a cumulative minimum would apply such a minimum to all options. If that is the case then we believe that participant elections should have no effect on the application of such minimums. If that is not the case, we do not believe that the cumulative minimum should be applied when there is a change in the interest credit option as this would effectively permit the cumulative minimum to become an annual minimum. It would also provide an incentive for participants with balances that are under the minimum to change their option solely for the purpose having the minimum applied to increase their balance.

On the other hand, it is not clear what to do if the minimum is not applied upon changing an election. The alternatives would seem to be to disregard the minimum, apply the minimum to the new option and maintain a no-cutback amount relative to the old option. As we believe that most plan designs would do this anyway, we would suggest having the minimum apply to the entire account, however the plan sponsor should be permitted instead to maintain anti-cutback amounts if they so desire. We also believe that the plan should be permitted to prohibit elections to change the interest crediting rate on existing balances to avoid this issue of how to address cumulative minimums.

We suggest that the plan document specify the option, from among the choices available to participants, on which the fixed interest credit upon plan termination would be based. It would seem administratively complex and unfair to determine the rate on an individual basis based on elections, as those elections were made with the expectation that they could be periodically changed. The crediting rates could be averaged over all of the alternatives or over the actual elections of the entire population but it seems preferable to us to have the plan specify the option. The regulations would have to make clear that there are no compliance issues associated with inserting this provision into the plan and that rules regarding the changing of options would apply to this as well. Alternately, perhaps the third segment rate could be used or the highest permitted fixed rate.

8. Plan Termination Provisions

We believe that the approach taken with regard to plan terminations in the proposed regulations is generally reasonable. We request clarification regarding whether past interest crediting rates that are included in the average include rates which are not permissible under the regulations once finalized. We observe that past rates could be impermissible under the final regulations for a variety of reasons ranging from the rates having fixed rates or minimums that exceed the amounts permitted to the rates being based on indices that are not part of the closed group of rates under the proposed regulations. If our advice in other portions of this letter is heeded (to expand the closed group and/or to increase the permissible margins so that all non-equity maximum rates are essentially viewed as equivalent) then we believe this situation would be substantially mitigated. We believe that past rates included in the average should be required to comply with the final regulations and that past rates which are noncompliant should be brought into compliance (solely for the purpose of determining the average) according to the process outlined above in this letter. However, we would suggest a numerical exception to this whereby any past rate would be permissible to be included in the average if it did not exceed the third segment rate for the stability period. Thus, for example, a fixed rate of 7% would need to be reduced to 5% (as currently proposed or 6% according to our suggestions) for purposes of the average, but only if and to the extent that 7% exceeded the third segment rate for the stability period (e.g., reduced to the greater of 5% [6%] and the third segment rate).

We do not believe that past impermissible rates should simply be replaced with the third segment rate. As previously discussed, many rates will not comply solely because of the closed group approach taken by the IRS. Many such rates will have expected values that are substantially less than the third segment rate and we don't believe it is appropriate to replace them with the third segment rate. This is one reason we are suggesting that noncompliant past rates be converted into compliant rates. In the alternative, noncompliant rates might be subject to the numerical exception (i.e., retained in the average if less than the third segment rate) articulated above.

To the extent that past rates are replaced with the third segment rate, we agree that minimums and maximums should attach to the third segment rate. We believe that adjustments (i.e., additions, subtractions or multiplications) should be examined on an expected (or historical) value basis. If the adjustment to the old rate is broadly representative of the difference in the expected values of the old rate and the third segment rate then the adjustment should be discarded. If the adjustment exacerbates any difference in expected values then it should be applied to the third segment rate.