Interest & Compensation Rate Components

[Andy the Actuary]

I recall a very old (circa late 1960s) actuarial exam study note [if you got a copy, please post] that discussed the components of interest and compensation. Interest consisted of pure interest (2%), risk (1/2%), and inflation. Compensation consisted of seniority (1/2%), merit (1%), and about 80% of the inflation. Thus, we would use 5% interest and 3 1/2 % salary scale; 8% interest and 6% salary scale.

It is clear that this classical model is as old and tired as my memory. Likely the risk component may comprise more of the interest and pure interest may be higher. Likewise, most of compensation is related to inflation. Thus, 8%/3% may be appropriate.

I drag you guys into this because PPA as usual poses a dilemma. What do you assume for compensation improvement when determining TNC versus what you use for determining the cushion amount. I could envision say for 2009, 0% for TNC, and 3% for determining the cushion amount.

Any thoughts?

[david rigby]

I go back to first principles. Use assumptions that are reasonable, and apply to the particular population. If that means a select and ultimate salary scale, so be it. I don't think it makes sense to use one scale for TNC and an entirely different scale for the cushion (and I doubt the IRS would approve).

Also, there is nothing wrong with using a 1% scale for plan X becuase it is in a depressed industry, but 5% for another sponsor. When setting salary scales, I don't have to apply "national inflation" if it does not affect the plan sponsor at the same time, or same rate, as the national average.

But I'm interested in other opinions also.

[tymesup]

Our system doesn't provide for different salary scales for TNC and cushion. I think you would have to have the same scale for the first year.

I suppose we could have a salary scale array, but that just raises other issues. Are salaries really geometric? Aren't they also a function of age? Service? Compensation? Ownership? Demographic group? Comp history?

[Andy the Actuary]

Our system doesn't provide for different salary scales for TNC and cushion. I think you would have to have the same scale for the first year.

I suppose we could have a salary scale array, but that just raises other issues. Are salaries really geometric? Aren't they also a function of age? Service? Compensation? Ownership? Demographic group? Comp history?

Or, we could use the Sarason tables (I once met Harry Sarason but that's another story).