This discussion pertains to Plans of fewer than 500 participants. In short, we're dealing with plan sponsors where there is no benefit expert HR or CFO and more often than not, you're dealing with the company owner.
Without regard to employing the full yield curve, there appear to be 10 possibilities in selecting funding segment rates -- with or without phase-in and using one of five look-back months. If phase-in is employed, then employer may elect (without IRS approval) to revoke the phase-in election in 2009. However, adopting the phase-in after 2008 requires IRS approval.
In establishing the interest rate assumption historically, the type of investments and long-term investment policy would be considered. Now, it appears we're into crystal-balling the bond market and all we can say is here's the effect for 2008 and here's the potential effect on 2009 if your decision regarding 2008 goes south.
However (if I've got this right and please comment if you see otherwise), this is more complication than sponsors of small Plans want to take the time for, if they could even understand it. Consequently, they will say "do it." As such, unless there is some compelling reason to do otherwise, I intend to recommend "phase-in" with fifth look back month. This seems to offer the most flexibility (5th month facilitates more reliable pre-plan year estimates). This means for most 2008 calendar year valuations, I would be proposing using (5.66%, 5.85%, and 6.03%). For this selection, only the look-back month appears to require an employer election.
There are two issues: (1) What should you being telling the plan sponsor as far as what his options are and their ramifications and (2) How does the employer make the election?
This is a damned-if-you-do and damned-if-you-don't situation. It seems as if you're almost greenmailing the client into going with your recommendation, which while not good, is what he wants. We do so much of this anyway without getting into "cya" exercises. The alternative is to paper the client to death describing the meaning of the election and the potential consequences (and why we're not using the full yield curve!). Since the phase-in is for two years, except in certain situations, I'm thinking about simply downplaying it. In short, it generally will not be a material decision.
I am not seeking universal agreement and would welcome comments as to approaches considered.