by Alvin D. Lurie

CASH BALANCE CONTROVERSY COSTLY TO ALL OF US BUT ESPECIALLY IBM

A heated controversy has been raging for a full year since a federal district judge in Illinois ruled that IBM's cash balance pension plan violated the age discrimination rules of ADEA and ERISA because of the manner in which it, like essentially every other plan in the cash balance format, accumulated the interest credits that are an inherent feature of such plans. The decision in the case, Cooper v. IBM, set off shock waves: first, in the pension community that has since gone into fast-freeze mode almost completely halting implementation of cash balance plans; second, in the Treasury that has halted a regulatory proposal initiated in 2002 that would have greatly facilitated the initiation of, or conversion of existing plans to, the cash balance design; and, finally, in the Congress that enacted one of its stranger pieces of legislation to squelch that regulatory initiative, by tacking onto an appropriations bill a prohibition against using any of the appropriation to implement the regulation, as an oblique way to insulate the judge's decision.

In the view of most pension specialists -- this writer included -- the opinion is a gross miscarriage of justice. Unless it is overturned on the appeal that is now certain to occur, following the recent partial settlement that would cap IBM's exposure at $1.4 billion should the Circuit Court uphold the lower court's age discrimination ruling, the decision would invalidate the cash balance formulae of essentially every cash balance plan in existence, estimated to number 11,000, covering over 7 million participants, representing 25 percent of all participants in defined benefit pension plans, with plan assets in the $600-700 billion range. But for the self-imposed moratorium in the pension community noted above, those numbers could grow geometrically, since many more employers were poised to migrate from traditional defined benefit plans to cash balance plans.

Repercussions from the IBM decision are not good for cash balance plans, or even for their participants, who will find that employers have dropped their plans (or worse, dropped employees or gone into bankruptcy paying claims in IBM copy-cat suits). Even if the case is reversed in a year or two after the appeals process has run its course, disenchantment with the cash balance format is likely to be long-lived among employers who are now living with the uncertainty. In a larger sense, this is not good for pension plans generally. Cash balance plans are a variety of defined benefit plans, but much different from the so-called traditional defined benefit plans that have long dominated the pension scene. Adoptions and continuances of the traditional variants have been in steep decline in recent years, dropping a staggering 72 percent since 1985 (from 114,500 plans to 32,500 in 2002), and certainly even further in the past year due to heightened concerns regarding the interest rate issue that attracted so much attention in this venue (see "The Old Switcheroo", Barron's 2/24/04) and elsewhere. At the very time of such shrinkage, cash balance plans have become extremely popular with employers of all sizes, but most especially large employers. Fully one-third of America's largest companies have turned to them.

It is fair to say that cash balance plans held, before the IBM decision, the promise to be the savior of the defined benefit tradition. Why does this matter? Defined benefit (or DB) plans have two distinct advantages over the defined contribution model, such as profit-sharing and 401(k) plans: a firm employer promise to deliver the stated benefit; and the guaranty of a government agency to back that promise, both of which have taken on additional sheen in the wake of the lack-luster, worrisome (and worse) performance of the stock market in the past several years. Cash balance (or CB) plans, as a subspecies of DB plans, share those characteristics, but have unique characteristics that provide even more attractive features for plan sponsors and participants. In this age of mobility, employees particularly like the portability of such plans, reinforced by their overcoming the disproportionate buildup of benefits late in an employee's working career that characterizes the traditional defined benefit plan; and mobility is further bolstered by the lump-sum payout of benefits whenever the employee separates from work, without reference to attainment of retirement age, a feature that is not usually found in traditional plans.

Such plans have an equally powerful appeal for employers, who have discovered to the shock of many of them that the promise to pay the backloaded benefits underwritten in their traditional defined benefit plans can result in unimaginable and unmanageable costs when the stock market slump combines with the deep interest rate decline (the so-called "perfect storm" of the recent past). The proof is in the pudding, evidenced by how rapidly the cash balance format has won approval among growing ranks of employers and employees alike. Then along came the IBM decision.

It is therefore important to know how the trial judge in IBM could have concluded that so many plan sponsors, to say nothing of the Treasury, got it so wrong, or rather whether -- as most commentators have stated, with the support of the Treasury's published positions on the issue (not referencing the IBM decision itself) -- it was the judge who got it wrong and is likely to be overturned. As you might suspect, the issue is not without its complications, primarily because of the existence of statute law that, if read literally and out of context, could lend some colorable support to the decision. But the obstacle to reliance on that language as an expression of Congressional intent is that it was written before the existence of the cash balance design and patently cannot be applied to the very fundamental differences of the cash balance plan from the traditional defined benefit plan for which the statute was crafted, without leading to absurd results that could not have been in the contemplation of the Congress.

The problem is not so much the complexity of the issues -- pension law is full of such -- as that cash balance plans are tested for age discrimination under a rule that neatly fits the traditional DB plan but is totally unsuited to the CB format, with the consequence that, arguably, CBs are unable to satisfy the rule if it is interpreted mechanically without regard to its unfitness. Specifically, the implicated rule requires a demonstration that the (annual) rate of accrual of plan benefits is not reduced because of attainment of any age. The words themselves, relatively simple on their face, admit of different interpretations, especially "rate of benefit accrual" and A reduction because of the attainment of any age.

The threshold questions are, what is a rate of accrual for this purpose and what elements entering into the building of a benefit are taken into account in measuring the decrease in accruals proscribed by the statute? Then, most importantly, if a progressive reduction in accruals, within the meaning of the statute were even conceded for argument purposes, is it prohibited even if not designed to effect age discrimination but merely occurs in observance of sound actuarial, economic and mathematical principles?

These interpretive questions do not arise in case of the traditional DB plan, where the benefits typically grow incrementally under a formula that adds discrete blocks of the ultimate retirement benefit for each year of service (such as a fixed percentage, say, 2 percent, of the participants compensation). Each year's segment thus becomes part of the participant's so-called accrued benefit; and the yearly rate of benefit accrual is readily ascertainable and typically equals or exceeds each of the prior years' benefit segments.

The CB formula for each yearly benefit segment is quite different. While it too is composed of a percentage of compensation (called a pay credit), there is also an interest credit, that is, an amount applied to the sum of accumulated prior years credits. The formula might be, for example, a pay credit of 5 percent plus an interest credit of 6 percent of all prior credits. The aggregate of these credits for all years at any given point in time becomes the notional account balance of a participant B notional, or hypothetical, because the participant does not really have an account on the plan's books, merely the promise of the plan (really of the plan sponsor) -- to pay the notional account balance on separation from service.

The interest credits, as the name implies, can be viewed as interest to be paid on the debt to the participant established by the plan, but payable only on the deferred date(s) specified in the plan. As with any interest amount, it is only payable for as long as the underlying debt is unpaid. Obviously, then, the aggregate number of interest payments stemming from any year's pay credit is strictly a function of the number of years remaining to the payment of the plan benefit, presumably the retirement date at the latest. Consequently, the farther the participant is away from that payment date (i.e., the younger, if judged by the distance from the normal retirement date), the greater the number of such interest credits the participant will accrue. That is not a matter of age discrimination but simply an aspect of the time-value-of-money principle; and, of course, the number of those interest credits will directly influence the benefit a participant will ultimately receive under the usual CB benefit formula. It is inconceivable that anyone could find anything objectionable about that.

However, another complication enters the picture, the so-called accrued benefit concept, because ERISA essentially requires benefits under DB plans to be expressed in terms of the annual benefit commencing at normal retirement age (typically 65, unless an earlier age is stipulated in the plan). This works fine for traditional DB plans, because that corresponds to the amount and time of payment of the benefit such plans are programmed to deliver. But CB plans are patterned on the so-called defined contribution model, where the accrued benefit is the balance in the employee=s account (although, as we have seen, it is a notional account in CB plans). There is a disconnect between the account balance and the accrued benefit at any given time, because of the interest credits, and, more specifically, because of how the IRS has prescribed the calculation of interest credits in determining the segment of accrued benefit attributable to any year's employment service. Instead of a simple application of the stated interest amount to the pay credit for the year (e.g., 6 percent of a $2500 pay credit, or $150), IRS requires the interest credit to be projected forward as an iterative addition of the interest amount for each of the years between the earning of the pay credit and the normal retirement age of each participant (in simplified terms, 15 years of projection for a 50 year old, 25 years of projection for a 40 year old). But this methodology is only prescribed for determining the accrued benefit; and, in fact, IRS has stated unequivocally that it is not applicable in ascertaining the accrual rate of the age discrimination rule.

Moreover, it has no effect on the actual benefit a participant will receive on separation before normal retirement age. The employee who resigns or is dismissed five years before retirement age will not get the interest credits that were projected in his or her accrued benefit for those five years. If it were otherwise, what a windfall that would provide for those who took a self-declared early retirement without loss of the portion of their accrued benefit that is composed of post-separation interest credits. The interest projection method is simply unsuited to the CB format, and is the source of the problem.

In any case, why should this accrued benefit quirk in the rules, as they relate to CB plans, have any bearing on whether a discriminatory reduction in the rate of accrual of benefits has occurred for purposes of the age discrimination statutes? Of course, the rate of accrual will be lower for older participants relative to younger participants B even for 25 year olds relative to 24 year olds B if the projection of interest credits to retirement age is built into the equation. A simple solution is to exclude the interest credits from calculation of the accrued benefit for purposes of testing for age discrimination under the "rate of accrual" methodology. That is exactly what the Treasury has proposed in those 2002 regulations that Congress deep-sixed, which would effectively neutralize the age discrimination issue by authorizing a special class of "eligible cash balance plans" that can use an alternative method for determining the rate of benefit accrual by reference to additions to the hypothetical account balance, and disregarding interest credits so long as interest is in fact accrued annually (i.e., not projected forward) at a rate unaffected by age.

Unless some such approach is taken, no cash balance plan can ever satisfy the age discrimination statute. If that were to be the law in this matter, then the law is, as Micawber would put it, for confusing economic precision for age discrimination.