Comment on FR Doc # N/A

1505 Gracewood Drive Greensboro, NC 27408 February 27, 2008 Internal Revenue Service http://www.regulations.gov Re: Proposed Regulation ??1.411(a)(13)-1 and 1.411(b)(5)-1 Gentlemen: Thank you for issuing Proposed Regulation ??1.411(a)(13)-1 and 1.411(b)(5)-1. The proposed regulations request a lot of comments concerning quite a wide range of issues, so this letter is long. Some of the comments were also requested a year ago in Notice 2007-6. I apologize for waiting until now to make comments. At your request, enclosed are comments, and they are organized into eleven sections, as follows: A. Meaning Apparently Reversed in ?1.411(a)(13)-1(d)(3)(ii) B. Conversion Protection Safe Harbor C. Constraining the Volatility of a Market Rate of Return D. Market Rate of Return Based on an Index E. Market Rate of Return with a Floor F. Adjusting a Market Rate of Return for Frequency of Compounding G. Bona Fide Market Rate of Return Already in Place H. Pension Equity Plans I. Floor-Offset Plans J. Conversion Amendment Definition in ?1.411(b)(5)-1(c)(4)(i) K. Lesser Market Rate of Return in ?1.411(b)(5)-1(d)(1) The sections listed above follow the order of the text of the proposed regulation, by and large. Except for sections A, J and K, the comments are responses to particular requests in the proposed regulations. Section C draws attention to the subtle distinction between arithmetic mean expected return and geometric mean expected return. This distinction is absolutely central to the issue of a volatile portfolio exceeding a market rate of return. Section E makes a detailed suggestion about a floor on an interest credit. Thank you for allowing the opportunity to submit comments. Any comments expressed here are my professional opinion and not necessarily the opinion of my employer. Faithfully yours, Thomas M. Zavist, FSA, EA A. Meaning Apparently Reversed in ?1.411(a)(13)-1(d)(3)(ii) Within the ?Explanation of Provisions? is a section entitled ?Section 411(a)(13): Special Vesting Rules for Applicable Defined Benefit Plans and Applicable Definitions.? Toward the end of the fifth paragraph is a sentence that ends with the clause, ?that are reasonably expected to result in a larger annual benefit at normal retirement age (or at commencement of benefits, if later) for the participant, when compared to a similarly situated, younger individual who is or could be a participant in the plan.? This sentence is saying that a pension plan that includes periodic adjustments that make a larger benefit for an older participant than for a younger participant has an effect similar to a cash balance plan. I disagree. The opposite is true. The sentence appears to include the words ?larger? and ?younger? in error. Perhaps the author intended to say ?smaller? instead of ?larger? or ?older? instead of ?younger? (but not both). Proposed Regulation ?1.411(a)(13)-1(d)(3)(ii) appears to have the same error. B. Conversion Protection Safe Harbor Further down within the ?Explanation of Provisions? is a section entitled ?B. Conversion Protection.? In the fourth paragraph, the IRS and the Treasury Department seek comments on ?another alternative means of satisfying the conversion requirements that would involve establishing an opening hypothetical account balance, but in limited situations would not require the subsequent comparison.? The problem here is designing a cash balance benefit after conversion that is at least equal to an existing accrued benefit in every optional form. For an optional form that does not require the ?417(e) interest rate for conversion, an interest credit and conversion factor from the cash balance account to the optional form can be designed relatively easily to be at least the existing accrued benefit, taking into account its early retirement reduction factors and its actuarial equivalence factors. In this case, the early retirement reduction and actuarial equivalence factors are all fixed, and the interest credit is at least a certain fixed rate. If this is the limited sort of case the IRS and Treasury have in mind, then existing rules will accommodate it, with no need for a safe harbor, except that a constant interest credit or being able to put a floor on the interest credit would be needed. The difficult case is an optional form that requires the ?417(e) interest rate, such as a lump sum. It is not obvious how to design a cash balance account without whipsaw that will always be at least equal to a lump sum benefit that can be arbitrarily large, depending on how low the ?417(e) interest rate drops. I suppose you could seed the cash balance account with an interest rate for conversion of 0%, but this might be too expensive for the typical plan sponsor. The discussion in the paragraph mentioned the case of a lump sum benefit that happens to be unavailable before conversion. In this case, it might make sense for the IRS and the Treasury Department not to require comparison to the accrued benefit converted at the ?417(e) rate. C. Constraining the Volatility of a Market Rate of Return Further down within the ?Explanation of Provisions? is a section entitled ?C. Market Rate of Return Limitation.? The twelfth paragraph requests comments ?on what other asset portfolios have sufficiently constrained volatility that they should be permitted to form the basis of a market rate of return for interest crediting under a statutory hybrid plan.? The key here is to distinguish between arithmetic mean expected return and geometric mean expected return. Practitioners typically model the future value of a portfolio as having a lognormal distribution. Letting ? denote the mean of the underlying normal distribution and ? denote its standard deviation, the arithmetic mean of the lognormal distribution is exp (? + ???), and its geometric mean is exp ?. The difference between the logarithms of these two quantities is ???, and as a rule of thumb, investors use half the square of the volatility (???) to approximate the difference between arithmetic mean return and geometric mean return. According to the Capital Asset Pricing Model, the quantity ? + ??? is a linear function of ? (e.g., ? + ??? = r + q?) so the quantity ? is a quadratic function of ? (i.e., ? = r + q? ? ???). Starting from zero volatility, as ? increases, ? increases until it reaches a maximum and then starts falling again. The graph of ? as a function of ? is a concave parabola. The graph of ? + ??? as a function of ? is a line. The yield curve is a graph of exp ? ? 1 as a function of duration, and duration is a monotone function of ?, so the yield curve has a unique maximum, and it looks something like a concave parabola. (Short term expectations also distort the yield curve at the lowest durations.) Bond investors are accustomed to looking at yields, which represent geometric mean expected return, so they are accustomed to looking at yield as a function of duration, which is a curve that looks something like a concave parabola. Stock investors are accustomed to the Capital Asset Pricing Model, which involves an instantaneous return that is the logarithm of arithmetic mean expected return. The graph of this instantaneous return as a function of risk (volatility) is a line. Both sorts of investors look at risk and return. Ultimately, there is only one asset model. Yet, practice is for one investor to look at geometric mean expected return and see something akin to a concave parabola, while another looks at arithmetic mean expected return and sees something like a line. Bond investors do not worry about increasing yield arbitrarily by increasing duration, because their yield curve has a unique maximum. Stock investors suppose you can increase return arbitrarily by increasing volatility, because their graph is a line with no maximum. Bond investors can add half the square of the volatility of each bond to their yield curve and change it from something akin to a concave parabola into something that looks like a line. Likewise, stock investors can graph geometric mean expected return instead of arithmetic mean expected return and turn their line into a concave parabola. There is only one asset model, and it applies to both bonds and stocks. Diversifying a portfolio is essentially equivalent to maximizing geometric mean expected return. (Some portfolios?such as pension investments?are tied to certain liabilities and to certain funding sources and an analysis of proper diversification should integrate their liabilities and funding sources.) For the purposes of asset allocation and diversification, the overall geometric mean expected return of the portfolio (as adjusted by other concerns) is the right measure to analyze and to maximize. The market will let an investor diversify optimally and thereby maximize geometric mean expected return. Alternatively, the market will let an investor be either ? more cautious than the optimum and thereby earn a smaller return or ? more aggressive than the optimum and thereby earn a higher arithmetic mean expected return but a lower geometric mean expected return. Whenever one investment strategy has a higher geometric mean expected return than another, it will outperform the other with a probability that rises over time and tends to 100% in the long run. Both the cautious investor and the aggressive investor, therefore, will underperform the optimal approach with probability that tends to 100% in the long run. The market permits all three styles of investment? optimal, cautious and aggressive. In the long run, however, the market rewards the optimal investor relative to both the cautious investor and the aggressive investor. The Pension Protection Act (PPA) of 2006 calls for an interest credit based on the market, and the market permits all three styles of investment?optimal, cautious and aggressive. The market lets an investor achieve an arbitrarily high arithmetic mean expected return by taking on arbitrarily high risk, but the market punishes the investor with a lower geometric mean expected return, as well as most likely a lower return (with probability that tends to 100% in the long run). The PPA contemplates a market interest credit and does not seem to contemplate limiting a hypothetical investor to any particular degree of caution. Pension plan sponsors tend toward an optimal investment, because the optimal investment is diversified and prudent. You ask what ?other asset portfolios have sufficiently constrained volatility that they should be permitted to form the basis of a market rate of return for interest crediting under a statutory hybrid plan.? All portfolios do. So long as an interest crediting rate can be achieved in the market, it will not exceed a market rate of return in the long run, no matter how aggressive it is. Geometric mean expected return is capped automatically at the return of a prudent, diversified portfolio. There is no reason for a regulatory limit on the choice of portfolio. There is no need for additional restriction prohibiting aggressive investment and there is no mandate in the law for imposing an additional restriction in a regulation. If you try to prohibit excessive risk in a regulation, it is not clear where to draw the line. Is a portfolio of 90% stocks and 10% at-the-money call options intrinsically too aggressive? What about 80% and 20%? Why limit the plan design? In the long run, the market will reward an investment that approximates a prudent, diversified portfolio. Should the government dictate whether a prudent portfolio must be 60% stocks and 40% bonds?or LDI with heavy emphasis on swaps?or any other particular asset allocation? In theory, there is a unique best way to invest assets in a given situation, which every investor constantly is trying to find. Investors may disagree, because they have different situations and different estimates of the optimal course of action, but they are all aiming for the same, perfect approach to investment. This optimal investment?a prudent, diversified portfolio of assets?arguably offers a market rate of return. Every other portfolio offers less return in the long run?but also a market rate of return. Later, the twelfth paragraph asks whether a regulated investment company described in ?851 is sufficiently diversified so as not to exceed a market rate of return. Yes, maximizing geometric mean expected return leads to diversification and leverage, so any failure to diversify will lower the geometric mean expected return and will not cause return to exceed a market rate of return. D. Market Rate of Return Based on an Index In the ?Explanation of Provisions? is a section entitled ?C. Market Rate of Return Limitation.? Also in the twelfth paragraph, ?Comments are requested on . . . whether it is appropriate to base an interest crediting rate on the value of an index.? Yes, an index is a good idea. An index may exclude dividends, so the increases in such an index may have to be increased by as much as 5% per annum for dividends. E. Market Rate of Return with a Floor In the ?Explanation of Provisions? is a section entitled ?C. Market Rate of Return Limitation.? The thirteenth paragraph asks whether an interest credit computed daily with a floor of 0% might greatly exceed a market rate of return. Yes, definitely it would. Applying a daily floor of 0% to an investment with a geometric mean expected annual return of 8% and a volatility of 10% will increase the geometric mean expected annual return to something in excess of 100% per annum. The fourteenth paragraph says that an equity index with an annual floor typically would exceed a market interest rate. Yes, I agree. An annual floor does not lead to as dramatic an excess return as a daily floor, but it exceeds a market rate of return. Also in the fourteenth paragraph, ?Comments are requested on what types of reductions to the variable rate would be appropriate in order to ensure that the effective interest crediting rate under these situations does not exceed a market rate of return.? First, certain types of interest credits need to be defined as permissible. These might include interest credits based on the sum of a particular Treasury bond yield plus a constant or the sum of a particular index or published corporate bond yield plus a constant or the return of an actual portfolio, such as the pension trust. Ideally, these will permit a look-back, so statistics for investment performance for one period can be used after a reasonable administrative delay for a later period. The permissible interest credits might also include a constant return, e.g., 5% per annum. (The fourteenth paragraph mentions floors of 3% or 4% per year. Why not 5%?) In general, the return on any actual asset, trust or fund ought to represent a market rate of return?even if lagged for a reasonable period of time. Likewise, the return on any aggregation of assets ought to represent a market rate of return, if the allocation is fixed before the return is known. Second, why not generally prohibit putting a floor on a market rate of return, except for certain safe harbors? Here is an example of a safe harbor, as well as an example of an impermissible floor. ? Safe Harbor?Permissible Floor. The interest credit for each period since inception is computed as the greater of a fixed quantity not to exceed 5% per annum or (1 + X) / (1 + Y) ? 1, where X is the total return compounded for the period and all prior periods, if any, since inception using a basis that meets the definition of a market rate of return and where Y is the total interest credit compounded for all prior periods, if any, since inception. This interest credit does not exceed a market rate of return. Suppose inception is January 1, 2009, the fixed quantity is 5%, the period is a year and the underlying return using a basis that meets the definition of a market rate of return is ?2% in 2009, 20% in 2010 and 10% in 2011. Suppose the interest credit for a year is the same for all participants, regardless when they enter the plan. o For 2009, X = ?2%, and Y = 0%, so the interest credit for 2009 is 5% (the greater of 5% or ?2%). o For 2010, X = 17.6%, and Y = 5%, so the interest credit for 2010 is 12% (the greater of 5% or 12%). o For 2011, X = 29.36%, and Y = 17.6%, so the interest credit for 2011 is 10% (the greater of 5% or 10%). ? Example?Impermissible 5% Floor. The interest credit in each period is computed as the greater of 5% per annum or a basis that meets the definition of a market rate of return. This interest credit exceeds a market rate of return. Suppose the period is a year and the underlying return using a basis that meets the definition of a market rate of return is ?2% in 2009, 20% in 2010 and 10% in 2011. o For 2009, the interest credit is 5% (the greater of 5% or ?2%). o For 2010, the interest credit is 20% (the greater of 5% or 20%). o For 2011, the interest credit is 10% (the greater of 5% or 10%). The safe harbor above bears some similarity to the safe harbor in ?1.401(a)(9)-6 A- 14(b)(3). The twelfth paragraph also asks whether an equity index which provides for preservation of capital might exceed a market rate of return. The answer to this question might depend on the manner in which capital is preserved. The safe harbor and example above address this. F. Adjusting a Market Rate of Return for Frequency of Compounding In the ?Explanation of Provisions? is a section entitled ?C. Market Rate of Return Limitation.? The last sentence of the fourteenth paragraph requests comments ?on whether regulations should establish reductions in these situations where the determination of whether the fixed or variable interest crediting rate is greater is made more frequently than annually.? No, reductions do not seem necessary for this. Suppose, for example, that an interest credit were based on an equity return for the year divided by 365 and compounded daily. Suppose the equity return is 10% in some year. The interest credit would be 10.515578% in an ordinary year and 10.545856% in a leap year. Admittedly, this process increases the interest credit above the underlying market return by just over half a percent. From the perspective of the plan sponsor, however, it is good to have a little leeway?the ability to exceed a market rate of return by an amount that is de minimis. You can squeeze a little extra interest credit by playing around with the crediting frequency and the compounding, but you cannot squeeze very much, and there does not seem to be a need to impose a regulatory cap on what small amount of wiggle room might be available. G. Bona Fide Market Rate of Return Already in Place In the ?Explanation of Provisions? is a section entitled ?C. Market Rate of Return Limitation.? The fifteenth paragraph cautions plan sponsors against ?adopting interest crediting rates other than those explicitly permitted in these proposed regulations.? There are a lot of cash balance plan out there already, and some of these have interest crediting rates that do not exceed a market rate of return from a layman?s point of view and yet do not meet the requirements of this proposed regulation. An example might be the actual investment return of the pension trust. It would be good if final regulations were to include a means by which a plan sponsor, who has a bona fide market rate of return as an interest credit, could somehow convert it to an interest crediting rate that complies with the final regulations without suffering an impermissible reduction to accrued benefits pursuant to Internal Revenue Code ?411(d)(6). In the section entitled ?Section 1107 of PPA ?06 and Code Section 411(d)(6)? at the end of the second paragraph ?Comments are requested on whether 411(d)(6) relief is or is not appropriate for any additional amendments related to section 701 of PPA ?06 or these proposed regulations.? Yes, definitely, such relief is appropriate. As indicated above, the relief might address an amendment changing the plan design from a bona fide attempt to meet the law to strict compliance with the regulation. H. Pension Equity Plans In the ?Explanation of Provisions? is a section entitled ?Pension Equity Plans (PEPs).? The first bullet point under the second paragraph asks whether a plan design with an interest credit that applies only after active participation ceases is properly treated as a pension equity plan. The same bullet point also asks if the design is properly treated as a cash balance plan. It is properly treated as a cash balance plan. An interest credit of 0% transforms a cash balance account into a lump sum accumulation, i.e., a PEP that is potentially more general than simply a percent of pay. The PEP, therefore, is a special case of the cash balance plan, and the PEP based on a percent of pay is a special case of the PEP. The PEP is a cash balance plan with an interest credit of 0%. The second bullet point asks whether two particular plan designs ?provide for a lower rate of accrual for additional years of service (because no interest is credited if service is continued).? Yes, they do provide for a lower rate of accrual, so long as the interest credit is considered part of the accrued benefit. The effect is similar to a subsidized early retirement reduction, and the accrued benefit should be limited so that it does not go down with increasing age or service. The second bullet point also asks if this issue can ?be avoided by treating the annual rate at which the normal retirement benefit accrues as declining with each additional year of service.? Yes, it can, with the same caveat as above that the decline should not cross over and become a reduction in accrued benefit on account of increasing age or service. The third bullet point asks how the back-loading rules apply to these designs. Traditionally, the back-loading rules have applied to annuities. It does not make a lot of difference if the back-loading rules were changed, in the case of a hybrid plan, to apply to a lump sum instead. Guidance is always helpful. I. Floor-Offset Plans In the section entitled ?Comments and Requests for Public Hearing? in the first bullet point under the third paragraph comments are requested on ?The application of the 3-year vesting requirement in section 411(a)(13)(B) to a plan that is not a statutory hybrid plan when the plan is part of a floor-offset arrangement with a plan that includes a lump-sum based benefit formula.? 3-year vesting does not seem necessary in this case, but it is not an onerous hardship. Guidance is always helpful. The second bullet point asks ?Whether guidance should be issued under section 411(b)(5) as to whether a characteristic is indirectly on account of age.? Guidance is always helpful. The point of a floor-offset arrangement is that the sum of two plans provides a benefit that is the greater of two arrangements. If each arrangement separately does not impermissibly decrease on account of age, why would the greater of the two impermissibly decrease on account of age? Arguably it would not, so the characteristic of the floor-offset plan would not be indirectly on account of age. The third bullet point asks about whether guidance is needed for a plan design in which the accumulated benefit is neither an annuity at normal retirement nor a hypothetical account nor an accumulated percent of pay. Guidance is always helpful, but there may not be very many pension plans other than these types. J. Conversion Amendment Definition in ?1.411(b)(5)-1(c)(4)(i) Proposed Regulation ?1.411(b)(5)-1(c)(4)(i) defines a conversion amendment to include any change to a pension plan that already has both hybrid and non-hybrid components whenever the non-hybrid component is reduced but the hybrid component is not frozen (for the next three years pursuant to ?1.411(b)(5)-1(c)(4)(v) (A)(2)). This definition makes it awkward to maintain a pension plan with both hybrid and non-hybrid components, because, to avoid a conversion amendment, any reduction to the non-hybrid component forces you to freeze the hybrid component. Some plan sponsors may find the definition of a conversion amendment less annoying if ?1.411(b)(5)-1(c)(4)(i)(B) were rephrased to say ? After the effective date of the amendment, benefit accruals under a statutory hybrid benefit formula under the plan are created or increased. Rephrasing this way would permit the sponsor of a plan that has both hybrid and non-hybrid components to reduce the rate of accrual of a non-hybrid component without changing the existing hybrid component or incurring a conversion. K. Lesser Market Rate of Return in ?1.411(b)(5)-1(d)(1) Proposed Regulation ?1.411(b)(5)-1(d)(1)(v) modifies ?1.411(b)(5)-1(d)(1)(iii). It permits an interest crediting rate that is always less than a market rate of return. What about an interest crediting rate that is sometimes less than a market rate of return and sometimes equal to a market rate of return? Why should it be prohibited? Why not delete ?1.411(b)(5)-1(d)(1)(v) and change ?equal to? in ?1.411 (b)(5)-1(d)(1)(iii) to ?less than or equal to??