Announcement: SOA congratulates the new FSAs for June 2019.

Does Process Really Matter In Limited-Scope Projects?

Does Process Really Matter In Limited–Scope Projects?

By David E.Weissner

It seems everyone is under increasing pressure to do more with less. Under tightening budgets and lower resource availability, limited–scope  reviews of issues are becoming increasingly frequent and partial solutions are implemented to address immediate needs. While a partial solution can buy time until resources become available for a full–scope review, it can also have significant unintended consequences if the process of the limited review does not include steps to validate the tentative conclusions against the principles and axioms of the broader context.

For example, consider conducting a limited review of a digital photo by focusing only on a few hundred pixels. The optics of that perspective are very different from that of the entire photo, and the viewer will walk away thinking that the photo is something that it is not. Had the viewer looked at the entire picture, he or she would understand the context of the area under examination and reach a conclusion based on and consistent with all the information.

The process that derives the result is every bit as im–portant as the result itself, even more so in a limited–scope review. The process must include checkpoints to test the results of the limited review against the established principles and axioms of the broader context. This conclusion may seem obvious, but as the following mini case–study of the proposed accounting rules for cash balance plans shows, it helps to remind ourselves of this point every now and again.

The Proposed Cash Balance Accounting Rules: A Limited–Scope Review that Missed (and Contradicted) the Broader Context

In 2003, the Financial Accounting Standards Board (FASB) was asked for guidance on the proper method for valuing a cash balance plan.

In undertaking the request, the FASB was very clear that this was a limited pension project with the objective of providing guidance for cash balance plans only. There have been clear indications for a few years that the FASB is unhappy with some of the principles underlying FAS 87, but they did not want to tackle a full–blown pension project at this time.  Limited–scope reviews are common for the FASB and very appropriate given the question asked, so long as they are grounded in the principles of the broader context.

Given the parameters laid out, the request seemed fairly straightforward. All the U.S. and international accounting standards published to date are very clear on the two types of plans: a plan under which the benefit paid is solely a result of the contributions paid by the employer and the earnings thereon is a defined contribution plan, while all other plans are defined benefit plans.

Said differently, if the employer takes a risk of any kind, be it mortality, investment or other risk, that plan is a defined benefit plan for the purposes of FAS 87 and IAS 19. For example, a non–qualified defined contribution plan (a plan that has no assets backing the accounts) is a defined benefit plan since the plan sponsor is taking all the investment risk, even though the plan looks identical to a defined contribution plan to participants.

Surprisingly, the FASB did not find the conclusion that cash balance plans are defined benefit plans as straightforward as most practitioners and undertook substantive debate on the topic. The meeting minutes of the discussions do not reference the existing literature of the broader context. Despite this lack of context setting, however, the FASB ultimately did conclude that cash balance plans are defined benefit plans. Having resolved this first debate, they explored liability valuation methods.

The actuarial profession had a chance to present to the committee and was unable to provide a singular recommendation, instead pointing out the pros and cons of four common methods.

The FASB continued its exploration and debate. From the meeting minutes, it seems clear that they were not looking at cash balance plan accounting in the defined benefit FAS 87 context, but as a different type of plan altogether—neither defined benefit nor defined contribution. The conclusion that cash balance plans were fundamentally different should have immediately triggered a hypothesis–testing step in the process. When results are unexpected, good process revalidates the conclusions by testing against existing principles and axioms to ensure the decision logic is valid or to determine whether a first principle needs revision; poor process forges ahead.

Many practitioners in the pension community were surprised when the FASB issued its initial conclusion. Boiled down, it says the four methods commonly being used are inappropriate for cash balance plans. It focuses on the notion that, since cash balance plans define account balances, the liability must be equal to the sum of the account balances.

In essence, the proposal suggests that since the benefits under a cash balance plan can be paid immediately, we should value the plan as if they will be paid immediately.  This is a problematic position because such treatment only applies to cash balance plans. All other pension plans that pay immediate lump sums can still discount the expected future benefit from the expected timing of the event.

This unfortunate outcome is a direct result of the limited–scope review that did not test the results against the principles and axiom of the current standard.  Indeed, the proposed rules contradict many of the key principles of FAS 87—reasonable assumptions about the timing and amounts of benefit payments, a reasonable actuarial method for allocating costs and the concept of discounting future cash flows. Viewed in the broader context, these rules will actually reduce financial comparability, not enhance it. Many practitioners were in a state of disbelief and thought that since the International Accounting Standards Board ("IASB") was looking at the same issue, perhaps cooler heads would prevail.

Then in a stunning turn, the IASB, perhaps in the name of convergence, perhaps not, followed with similar faulty logic for plans that are similar to cash balance plans. More disturbing still, the IASB, the mouthpiece for principles–based accounting had just proposed more rules that directly contradicted the principles and axioms of their current standard.

The FASB received several unsolicited comment letters from practitioners and users that hoped to demonstrate the contradictions of the initial proposal. The IASB also received solicited feedback on its proposal. At the time of this writing in mid–October, these letters had clearly begun to have the desired effect opening up discussions to the broader context.

In another shocking turn however, the FASB hastily decided that they would merely expand the initial conclusions from a cash balance only viewpoint to a project that would cover all pension plans that paid lump sums, instead of reviewing the project in the broader framework. As will be demonstrated, this new proposal creates new logical contradictions within the broader context.

What Went Wrong?

By limiting the scope of the project, the FASB purposely focused on a very narrow issue, but failed to revalidate results in the broader context. In hindsight, this was the equivalent of looking at a few hundred pixels of a five–mega–pixel digital photo—you get a distorted and confusing view. Once committed to only looking closely at one small area, it is very easy to forget how it fits into the big picture—in this case, the principles underlying FAS 87. When the FASB was then faced with a list of logical contradictions, it expanded the scope just enough to cover those contradictions, but not a pixel more.

In deciding to undertake a limited–scope review, the FASB should have begun by testing the hypothesis that cash balance plans are materially different from other defined benefit plans. Had they taken the step of substituting demonstrations for impressions, it is highly unlikely that the initial proposal would be so contradictory. The sidebar demonstrates that cash balance plans do not accrue or pay substantively different benefits from some traditional pension plans.

When the FASB in mid–October finally realized that cash balance plans are not different, they changed the focus of the project from a cash balance project, to a project covering plans that pay lump sums. Unfortunately, they are about to realize via yet another round of unsolicited comment letters that will substitute demonstrations for impressions that they won?t be able to stop there. The worst outcome of the new round of letters would be that the FASB then tries to do yet another incremental fix; the best outcome would be that the FASB realizes the only way to deal with the level of complexity that is inherent in pension plans is to open up a full–scale review.

By failing to test hypotheses and results against existing principles and axioms, the FASB evaluated cash balance plans as well as plans that pay lump sums, without the necessary context. Without such context, things may look very different. For example, one of the first things people notice when looking at cash balance plan liabilities is an optical problem in the results: due to the time value of money, the liability held is often less than the sum of the account balances. Does this happen in traditional plans that pay lump sums? Yes, but since there are no account balances to compare against the liabilities, there is no optics problem.

This poor optical result happens because cash balance plans typically credit earnings on account balances based on yields on U.S. Treasuries (perhaps with a specified margin), whereas the discount rate is typically based on the corporate bond market. By projecting a current balance out to the date of withdrawal with a lower rate, and discounting it back to today with a higher rate, the liability is typically lower than the account balance itself.

The initial proposal eliminates this optical problem by effectively eliminating any projections. This proposal is equivalent to assuming every participant terminates today and takes their cash balance immediately. While the proposal does solve the problem of reconciling the account balances with a smaller liability figure, 100 percent immediate termination of an entire workforce is clearly an unreasonable assumption for an ongoing plan.

These same problems exist in the newest (mid–October) thinking from the FASB for the project on accounting for pension plans that pay lump sums. For example, a plan does not have to offer a participant a lump sum to settle the plan?s obligation. That settlement can be achieved quite effectively via purchasing an individual or group annuity product from the plan?s assets and relieving the company of financial commitments. These types of settlements are usually special one–time operations of the plan and are currently accounted for under FAS 88, which is currently not part of the scope of the project. Thus, the project must now expand to cover the axioms and principles of FAS 88 to be cohesive.

The original draft of this paper cited the positives of how unsolicited comment letters led to promising signs that the FASB is reconsidering its cash balance proposal and how hopefully their next steps will test and validate future proposals against the current principles and axioms. With the latest thinking from the FASB issued in mid–October it is more likely than not that this process is going to be a long, drawn–out, frustrating and iterative affair for plan sponsors who will not be able to get a firm handle on how their future accounting obligations will look.

All of these contradictions and iterations are avoidable by redefining the process to look at the issues against the axioms and principles of the current standards. This process may lead FASB to conclude that some principles need updating. For example, perhaps discounting for the time value of money is inappropriate, or perhaps the discount rate guidance needs revising. In these cases, a comprehensive pension accounting review may be the only appropriate course of action. The final conclusions may be unpopular and startling, but they will be based on and consistent with a single cohesive set of principles and axioms.

What Lessons Can We Take Away From This?

The FASB is facing the same pressures we all are in trying to accomplish as much as possible with the limited resources available. In addition, the FASB is beginning a long transformation process that is akin to a corporate reorganization—convergence, transparency and moving to principles–based accounting are the goals. These changes will come more slowly than many would like and will be painful for some users of the standards. The FASB has a whole host of other issues on their plates besides pension accounting and their time and resources are very limited.

Notwithstanding limited resources, in nearly all aspects of accounting, the issues are going to be complex and it may be difficult for a single set of rules to cover all eventualities. When engaging in limited–scope reviews, especially in the current rules–based accounting systems, the FASB could improve their product by testing their tentative conclusions against the principles underlying the existing standard prior to release, in this case, what is pension accounting trying to achieve? How did it get to its current state? Why were those decisions made at the time and what has changed? How does the current issue fit in the global context of the standards and all other accounting standards?

We can all learn a good process lesson from this limited–scope project. Limited–scope projects are extremely challenging in the complex environments that companies are facing every day. This case review should remind all of us of the importance of validating tentative solutions against the existing principles and axioms and in the broader context in all situations, including limited–scope reviews.

David E. Weissner, FSA, is a consulting actuary for Mercer Human Resource Consulting. He can be contacted atDavid.Weissner@mercer.com

The views presented in this article are those of the author and do not represent the views of his employer.

Plan Architecture And Financial Comparability

There are many types of plan architectures that produce different accrual rates and benefit streams: final average pay, career average, dollars times years of service and cash balance plans are a few examples. The common result of all of these architectures is that they define how the benefit is earned each year and how the benefit is ultimately paid to the participant, either in lump sum form or in an annuity (or a combination thereof).

An interesting result is that each type of plan can be expressed as a function of other plan types. The resulting functions may not be easy to communicate to participants, but they are mathematically equivalent. For example, consider the following two plan designs:

  • Career Average Plan
    • 1 percent of career average pay times service, payable at age 65
    • Actuarial equivalence of 5 percent, no pre–retirement mortality. Post–retirement mortality is 5 percent GAM83 Unisex
    • Annual past service updates granted for any year inflation exceeds 5 percent. Update amount is prior year?s benefit times excess of CPI over 5 percent
    • Benefit payable immediately on termination in an actuarially equivalent lump sum
  • Variable Interest Cash Balance Plan(that has a tiered pay credit approach based on age)
    • Pay credit at each age is based on a table. The table is constructed from the following formula: 1.6753 percent of pay at age 25 * 1.05ˆ (Current Age–25)
    • Interest credits: greater of CPI or 5 percent per year
    • Benefit payable immediately on termination in a lump sum
    • These two plans produce identical accrual patterns and benefit streams under all conditions. Under the proposed accounting rules, a lower liability would be held under Plan A than Plan B merely because of the plan architecture.

Of course, these are very simplified examples. In real life, the cash balance plan would not have a tier at every age, although every five years is not uncommon. Variable interest crediting rates, early retirement provisions, actuarial equivalences, country specific regulations and other variables would make the exercise of designing a mirror image plan with a different architecture a bit more challenging.

Balancing out the challenge is the hundreds if not thousands of practitioners that will be working on the mirroring and the fact that plan sponsors will likely accept a "close enough"approach to mimicking the prior cash balance formula while creating a plan that is easier to communicate and understand.

The implications of the current proposal are far–reaching throughout any company that sponsors a cash balance defined benefit plan. First, most reasonable people would find it illogical that two cash flows using the same assumption would produce different liability results solely because of how the formula that generates those cash flows is expressed. Plan sponsors will be under tremendous pressure to modify the architecture of the plans to get the most favorable accounting. The duality of the accounting results demonstrated above will in turn reduce the comparability of pension footnotes. As the change–over occurs, significant expenses will be incurred in the design, communication and documentation of the new plans. Finally, employees are likely to see more complicated plan designs, causing even more confusion into how they actually earn their benefits.