De-risking activity: Impact of interest rate and mortality assumption changes

We recently published two articles, Accounting impact of changes in interest rate and mortality assumptions and Funding impact of changes in interest rate and mortality assumptions. Since we wrote those articles the Fed Open Market Committee met (in September) and decided not to raise the Fed funds rate, although reports are the committee is still mulling an increase in December.

In this article we are going to briefly discuss the effect of changes in mortality and interest rate assumptions on de-risking transactions. While de-risking covers a lot of things, we are only going to focus on its effect on de-risking transactions that involve paying out lump sums to terminated vested participants.

We’re going to keep this article short, and there will be no extensive mathematical analysis, because the basic point is pretty simple: If, as seems likely, for accounting sponsors adopt the new Society of Actuaries mortality tables in 2015 or 2016, but IRS does not adopt the new tables for purposes of determining lump sums (that is, for purposes of determining the amount of cash that must be paid out) until 2017, then, for any given lump sum payment in 2016, sponsors will be able to decrease liabilities on their balance sheet by a greater amount than the plan actually pays out in cash.

Background

A de-risking transaction is interest rate- and mortality assumption-dependent. That’s because generally (and with the exception of cash balance benefits) a DB liability is defined, and carried on the plan’s books, as an annuity benefit. Thus, you must use interest rate and mortality assumptions to determine (1) the lump sum (present) value to be paid out by the plan and (2) the amount of the associated liability that is considered ‘settled’ and subtracted from the sponsor’s balance sheet.

We assume that sponsors that intend to do a de-risking transaction in 2015 have their plans in place (if they have not already proceeded with the transaction). We are going to frame the analysis of de-risking transactions by considering the cost/benefit of such a transaction in 2016 vs. 2017. We don’t think it’s very useful to speculate about interest rates that far into the future. We do, however, have a pretty good idea of what is going to happen to mortality assumptions and when it’s going to happen, and our primary focus will be on the effect of new mortality assumptions on that de-risking decision.

Effect on the de-risking decision of the adoption of new mortality tables

At the end of 2014 the Society of Actuaries finalized new mortality tables for private DB plans (see our article Society of Actuaries releases updated mortality tables). DB plan sponsors will want to consult their actuary as to the application of the new tables to their plan.

Adoption of the new mortality assumptions will increase liabilities and lump sum values: longer lives = bigger lump sums. The following table projects (based on estimates) the increase in annuity values for minimum funding purposes that would result from the adoption of the RP-2014/MP-2014 mortality tables/improvement scale by IRS, assuming that prior to adoption the plan was using RP-2000 base tables and mortality projection Scales AA.

Table 1 — Increase in liabilities at different ages resulting from adoption of new mortality assumptions

Age

Increase

Males

25

9.8%

35

7.9%

45

5.4%

55

3.6%

65

4.5%

75

9.8%

85

16.9%

Females

25

11.8%

35

10.3%

45

8.3%

55

6.6%

65

5.8%

75

8.0%

85

10.7%

Source: Society of Actuaries RPEC Response to Comments on RP-2014 Mortality Tables Exposure Draft

For most plans, we estimate the impact of updating mortality is to increase both GAAP liabilities and lump sum benefits by 8%-10% (this does not apply to cash balance benefits, which are unaffected by mortality.)

Many, perhaps most, pension sponsors will have updated their GAAP mortality assumption by 2016, On the other hand, it seems clear that IRS will not be adopting new mortality tables for lump sum calculations until 2017 (see our article 2016 mortality tables published by IRS).

Accounting/cash mortality assumption mismatch for 2016

So, for 2016 at least, there will be a mismatch for many employers. Lump sums paid out will, for accounting purposes, be valued using the new mortality tables, but the amount of cash paid out will be determined using the old tables. The result: the amount of liabilities written off will be greater than the amount of the lump sums paid out.

In short, a plan may be able to reduce its GAAP liability by $11 million during 2016 by paying out $10 million in lump sum benefits. In 2017, those same lump sum benefits will cost $11 million if interest rates don’t change.

This example significantly oversimplifies. The exact amount of the 2016 cash/accounting mismatch will depend on the application of the new tables to the plan’s population and whether the sponsor, or IRS, makes any modifications to RP-2014.

Interest rates, briefly

Valuation interest rates, like mortality, get different treatment under cash and accounting rules. The amount of a lump sum to be paid out (the cash cost to the plan) is determined based on PPA ‘spot’ first, second and third segment rates for a designated month. Many sponsors set the lump sum rate at the beginning of the year, based on prior year November interest rates, so that participants will know what rate will be used to calculate their lump sum for the entire year. For those sponsors, the lump sum interest rate for 2016 will be pegged at November 2015 spot rates.

For accounting (that is, for purposes of determining the amount of liabilities that will be subtracted from the balance sheet when a lump sum is paid out), the lump sum valuation interest rate is typically set based on prior-year December 31 spot rates. So, e.g., for accounting purposes the lump sum interest rate for 2016 is based on the December 31, 2015 rates.

With respect to de-risking and interest rates we only want to make three very general observations.

First, if interest rates go up across the board by the end of the year, that will generally make valuations smaller, de-risking in 2016 less expensive, in terms of cash, and (correspondingly) the balance sheet impact smaller.

Second, it is possible to have a cash/accounting interest rate mismatch, particularly when interest rates are moving at the end of the year. In a context of decreasing interest rates, where year-end rates are lower than November rates, sponsors may find themselves in a situation where they ‘make money’ on a lump sum – the amount of liability that is written off is greater than the amount that is paid out, because year-end (accounting) interest rates are lower than November (cash) rates. This is another instance of the cash/accounting mismatch discussed with respect to 2016 vs. 2017 mortality assumptions.

Where – as some predict – interest rates are increasing during the remainder of 2015, year-end rates may be higher than November rates. In that situation (and disregarding the mortality assumption effects discussed above), sponsors that peg plan lump sum payout calculations to prior year November rates may ‘lose money’ on 2016 lump sums – the amount of liability that is written off may be smaller than the amount that is paid out, because year-end (accounting) interest rates are higher than November (cash) rates.

Third, as we discussed in our first article (Accounting impact of changes in interest rate and mortality assumptions), some believe that a Fed increase in short-term rates may result in or at least coincide with a decrease in long-term rates. The relationship between Fed interest rate policy and the valuation yield curve is not always linear.

* * *

We will continue to follow these issues.