ABO and CL for backloaded cash balance plans

[Guest Mike Melnick]

A cash balance has a higher interest credit for active employees than for former employees. In other words, it is partially backloaded. (The Contribution Credits are a percent of salary).

Both for purposes of measuring the ABO under FASB #87, and the current liability for ERISA purposes, in my opinion, it is appropriate to roll forward the balance at the lower rate (i.e the rate for inactives). Basically my reasoning is that the additional Interest Credit, like future Contribution Credits, has not yet been earned by the employee, who must perform future service for the employer before he is entitled to the additional Interest Credit. Therefore the additional Interest Credit should not be included in valuing the accrued benefit.

I have heard other actuaries express opinion on both sides of this question. I am curious about what interpretations other people are using, and whether the FASB or IRS has commented on the question.

[Blinky the 3-eyed Fish]

To the degree the actives are assumed to remain active, I would roll forward the interest at the active interest rate. If there are pre-retirement decrements, then it would seem the inactive interest rate would kick it only a participant is decremented out.

There's one opinion.

[Guest RSNOW]

Add my vote to Blinkys. I believe FAS 87 states something to the extent that your actuarial assumptions are supposed to be consistent with the valuation assumptions. On your normal actuarial valuation you are either assuming no pre-retirement decrements, or one or more decrements, either way most participants will still be assumed to remain employed until retirement. I think it would be difficult to justify a significantly different assumption on employee turnover for FASB (i.e., NO assumption of continued employment) which is inconsistent with the actuarial valuation assumptions.

[Guest Mike Melnick]

Thanks to both of the above for the feedback. I would like to present one additonal piece of information to the discussion: the Interest Credit is defined completely in the plan, without any reference to an economic index.

In this context, it is not a question of choosing actuarial assumptions. It is a question of interpreting plan provision. The choice of what Interest Credit to use in rolling forward the balance has nothing to do with making an assumption about economic conditions.

Looking at Paragraph 18 of FASB #87, the ABO is defined as being based only on service rendered prior to the measurement date, and that it represents the obligation the employer would have "if the plan were discontinued". Based on that, it appears that it is a closer fit to the Statement to use the inactive rate.

[MGB]

RSNOW: There is absolutely no connection from SFAS 87 to what you are doing in a funding valuation other than Actuarial Standards of Practice being applicable to both. Acceptable practices in funding may or may not meet the best estimate requirement of SFAS 87.

I initially felt only the inactive rate would be used for ABO, but haven't completely accepted that now. I think Blinky's description is correct for PBO.

The problem with the ABO is the statement in paragraph 19 that in non-pay-related plans, the ABO must equal the PBO. The decision by FASB's EITF two weeks ago was that a cash balance plan is to be treated as a non-pay-related plan.

All of this discussion is theoretical and temporary. Before the end of the year, the FASB will pronounce that the interest crediting rate is the discount rate, so no matter what you do in projections, the ABO and PBO will equal the account balance. (You project forward and discount back at the same rate.) The original observation by Jules Cassel at the EITF on this issue was too narrow for their liking (it was only focused on market-based crediting rates). That is why they delayed its implementation. They are expanding the concept further and I expect the final result to be ABO=PBO=Account balance for all plans once they are through (with additional implications for non-cash balance plans).

Of course, there is the bigger issue in this case: How can you possibly pass 411 accrual rules with this backloading?

[Guest Mike Melnick]

MGB: Thanks for the updates about the EITF. It does appear that the this entire debate may become irrelevant for future years.

Compliance with the anti-backloading rules of 411 is, of course, a very important constraint on the design of cash balance plans. Because there is some inherent "front loading" in the cash balance plans (when the accruals are expressed as deferred monthly benefit), there is some room for backloading. Successfully demonstrating compliance (generally based on the 133% rule) dependes on (1) the actual spread between the inactive and active rate, and (2) the actual level of Contribution Credits.

[Mike Preston]

Mike, would you mind giving a small numeric example of a set of design criteria which would satisfy 411 and then how a modification would result in a design which fails 411? If you don't have time, don't bother, but if you have an example handy would you share?

[Guest Mike Melnick]

In reply to Mike Preston, I have giiven an example below. Consider these plan specifications:

Contribution Credit = 5% of pay

Active Rate is 6.5%

Inactive rate is 4.0%.

Factor to convert a cash balance at age 65 to a monthlly annuity = 115.

Suppose participant enters plan at age 20, and earns 30,000.

Annual contribution credit (with Interest Credits to end of year) is

30000 * .05 * 1.065^.5 = 1547.98

Note that the accrued deferred monthly benefit (based on the existing cash balance) is calculated by rolling forward the accumulated cash balance at 4% (the inactive rate) to age 65 and dividing by 115. The additional 2.5% interest credit (applicable for actives) can then be thought of as pre-retirement COLA adjustment to the accrued benefit.

To measure the accrual each year, consider it consisting of 2 pieces. The first piece is the 1547.98 annual Contribution Credit rolled forward at 4% to age 65 and divided by 115 (this is the additional deferred monthly ben result from the current Contribution Credit). This piece shrinks each year, resulting in a kind of frontloading, which helps with the test. The second piece of the annual accrual is the 2.5% (the spread between the active and inactive rate) multiplied by the deferred accrued mo ben (based on the accumulated cash balance at the beginning of the year). This is the backloaded piece.

In this example, at age 35, the accumulated cash balance would be 37,434. (Note this amount includes all prior "COLA's". In other words, the Contribution Credits have been rolled forward to age 35 at 6.5%.) The 37434 converts to an accrued monthly benefit of 1055.75 at age 65 (1055.75=37434*1.04^30/115).

To measure the accrual during age 35, we now look at both pieces. The first piece of the new accrual is based on the Contribution Credit, and is 41.98 (41.98=1547.98*1.04^29/115). The second piece is the "COLA" adjustment to the accrued benefit, 26.39 = 1055.75*.025). The total accrual during age 35 is 68.37=41.98 + 26.39.

If you repeat this exercise to measure the accrual during age 64, you would get 94.08 = 13.46 + 80.62

Notice the COLA piece has grown much larger, while the Contribution piece has shrunk. However, he net result is a ratio of 138%= 94.08/68.37, which exceeds 133%, so the plan fails.

However, if you slightly increase the inactive rate to 4.25%, and recalculate all the amounts, the test ratio drops to 122%, and the plan passes (at least for these two ages). The actual numbers are as follows: (13.46+72.23)/(45.01+25.53) = 86.12/70.54 = 122%