Cash-Balance Plans

[Guest RSNOW]

I understand you can cross-tested a cash-balance plan by testing it on a benefits basis (subject to gateway req) . If true, if your plan's normal form of benefit is a lump sum and your actuarial equivalence in the plan for all optional forms of benefit use a standard mortality table & interest rate (7.5-8.5%), would you typically have an MVAR that is greater than the NAR ? Is the annual contribution (pay credit), using current year testing, simply projected to testing age at a 7.5-8.5% interest and converted to benefits using standard mortality table ? or does the 417(e) interest credits (if used) get included in the accrual and either result in a MVAR higher than NAR or generate a larger NAR than what I'd get simply projecting pay credit at 7.5-8.5% and dividing by annuity factor and then compensation. Any input on methodology would be appreciated.

[MGB]

You do not "cross-test" a cash balance plan on a benefits basis. The cash balance plan is already on a benefits basis. It is the annuity benefit defined by the plan that you use in testing. No cross-testing methodology applies here. You need to calculate the annuity benefits under the plan's forumulas (there should be some way to convert the cash balance into an annuity benefit at NRA) -- you should not be using 7.5% to 8.5% projections of the cash balance from the cross-testing rules.

[Guest RSNOW]

My original post may be incorrect. My question comes from a outline from a recent talk from a NIPA conference where it stated that cash-balance plans can be cross-tested. Perhaps I erroneously assumed it was on a benefits basis, but maybe the cross-testing is converting the DB benefit to an DC allocation and then general testing it as a DC like allocation. Is this allowed ? If so do you first calc the benefit under terms of the DB plan (using conversion factors) and then convert the benefit to an equivalent allocation (present value) using standard mortality table and interest for testing ?

[MGB]

Yes, you can cross-test back to a contribution basis as you describe. But, if your HCEs are older than the NHCEs, this won't help you much.

I am guessing that they are describing the conversion into annuity benefits under the plan as being similar or equivalent to cross-testing on a benefits basis, although the difference in interest projections produce very different results.

[IRC401]

Any idea how actuarial firms test cash balance plans with "self-directed" investments? Thank you.

[MGB]

It doesn't make any difference what the investment is. There needs to be a procedure written into the plan to convert the account balance into a deferred annuity (note 1). That deferred annuity is what is tested. There is no need to make any assumptions about the investment returns unless the aforementioned procedure makes reference to the latest yield on the balance at the time of the conversion.

Note 1: Having said that, I realize there are plans without such language and I don't know how they get by this, not only for nondiscrimination testing, but for a whole host of other issues.

[mbozek]

Bank of America adopted a self directed CB plan in 1998 with 9 investment options.

[MGB]

Just to clarify:

B of A's plan (along with other "self-directed" plans) does not have individual selection of investments. What they are selecting is the fund(s) that will produce the interest credit(s) on their account. The plan itself is not invested in these funds.

For example, I choose B of A's bond fund for my entire account. If that bond fund increases 10% this year, I get 10% interest credit on my account. However, the plan may have no investments in that bond fund (or, for hedging purposes, it may have investments in it).

The funds offered by B of A are the same ones that people had invested under their 401(k). When they went to a self-directed cash balance plan, the 401(k) plan was rolled over into the cash balance plan. The reasoning was that B of A felt that individuals were conservatively invested and that by having the trust invest in more aggressive investments, B of A gains from the arbitrage over time.

[mbozek]

B of A also got a one time increase in the corporate balance sheet from the infusion of 401(k) assets into the CB plan. From the participants viewpoint it doesnt matter whether their contributions are actually invested in the funds since the accrued benefit cannot decline. From the plans viewpoint if the participants are taking no risk, e.g., mm or stable value funds then the plan can invest the funds in equities and keep the difference in return or hedge the risk.

[Guest Harry O]

O.K. . . . I'll bite! Two questions:

1. Why can't the accrued benefit decline? If the notionally invested accounts decline in value, the accrued benefit declines as the result of the plan formula but not as the result of a prohibited amendment under section 411(d)(6).

There is the provision of the IRC that says that the normal retirement benefit can never be less than the dollar amount of any early retirement benefit but these plans (as I understand it) define normal retirement as 5 years of service.

I know the IRS has made noises from time to time that PIA offset plans cannot have a decrease in the accrued benefit from year-to-year if the offset grows faster than than the gross benefit, but this is a controversial and untested theory as I recall.

2. What do you mean by a one-time increase in the corporate balance sheet? Increase in assets? In any event, wouldn't every $1 of assets transferred to the pension plan bring with it an offsetting $1 of liabilities? Maybe there is some accounting gimmick at work here that is not obvious??? I can see an income statement benefit going forward if the amount earned on the transferred assets exceeds the amount of interest promised to employees on their transferred funds under the cash balance plan. But I don't follow the one-time balance sheet benefit . . .

[MGB]

On question 2:

Accounting for cash balance plans is very controversial and is applied in numerous different ways, depending on the actuary and auditor. Under many approaches, the liability is not the same as the account balance. So, if the liability is less than the assets that are transferred, there is an immediate gain. That is a point-in-time gain. It would be spread over average future service typically. However, if the cash balance plan at the time of the transfer (one already existed, they just added to it with the 401(k) transfers) were underfunded, the gain would be an immediate hit to the balance sheet, offsetting any previously recorded additional minimum liability. But, it would be hard to believe their plan was underfunded in 1998.

In a speech I gave last week, I also called cash balance gains as being mostly derived from "accounting gimmicks" (I was refering to Delta's announcement that they would save $500 million over the next few years even though they grandfathered all workers over 50 into the final pay plan - that doesn't add up; give younger workers more benefit and older workers the same benefit -- where does the savings come from other than accounting gimmicks? As vice-chairman of the Academy's Pension Accounting Committee, I take these public announcements about pension's affects on bottom lines very seriously. Note that the Emerging Issues Task Force of the FASB is addressing cash balance accounting in its meeting tomorrow at our committee's request.).

Then there is also the annual gain to the bottom line in the arbitrage between expected return on assets (high) and the increase in the liability (low), which you already recognized.

[SoCalActuary]

The NIPA conference gave an introduction to cash balance plans, and alluded to the possibilities available. A couple of points to follow up:

1. Cash balance plans are DB plans, and the hypothetical balance must be converted using the plan document's rules into an equivalent annuity benefit. That benefit is tested for TH minimums, 415 maximums, and EBAR's under 401(a)(4).

2. An annuity benefit can be cross-tested as a DC test. Since cash balance plans use lower interest rates for benefit conversion than the interest rates in 401(a)(4) testing, this can be advantageous to younger HCE's.

We use cash balance/401(k) combinations to get high flexibility of contribution rates, and test both plans together when it helps.