RP-2000 Mortality Table

[Guest sdolce]

Is the RP-2000 Table automatically to be used for purposes of 412(l),or is the Service required to make it official by regulation,ruling,etc.?

[david rigby]

Neither. The SOA taskforce did an exceptional job in looking at basic mortality data and has submitted this table to the actuarial profession and to the IRS with recommendation that it be considered as a replacement for the 1983 GAM table. That table is currently used for calculations under IRC 412(l) current liability and for determining PBGC variable premium liability.

The RP-2000 table is really multiple tables, with strong discussion and documentation illustrating the mortality variations by gender and by "collar" Although I have a personal concern over how the committee defined collar, it is clear that there is more than trivial differences between white collar and blue collar.

The ball is in the IRS's court. They could do nothing, adopt as recommended, adopt with some modification, adopt with some phase-in, etc. My guess is that they will not adopt anything right away, but will adopt with a phase-in. Because there is an entrenched mindset at all government levels that "one size fits all", I doubt they will take the advice of the taskforce in allowing different mortality tables by collar.

[david rigby]

BTW, the report can be found at this site. (PDF, about 110 pages.)

http://www.soa.org/research/rp2000.html

[david rigby]

This table, and its development, has hit the news lately.

http://www.actuary.org/pdf/pension/nyt_050603.pdf

http://www.plansponsor.com/pi_type10/?RECORD_ID=20596

http://www.actuary.org/pdf/pension/portman_050603.pdf

[Guest prp]

Is the RP-2000 Mortality Table also being considered for use in the calculation of lump sum distributions of Defined Benefit plans (if so, when?) or will the GAR 94 table continue to be used? Also, how does one calculate life expectancy for a person of a given age using the GAR 94 table, or is there a website that provides that information? Thanks.

[david rigby]

IRC section 417(e)(3) provides for the “applicable mortality table” and “applicable interest rate” used for calculating minimum lump sums. The current such mortality table is the unisex version of the GAR94 table (with projection) as the IRS has given in Revenue Ruling 2001-62.

Under current statute and regs., the mortality table is defined as “based on the prevailing commissioners' standard table (described in section 807(d)(5)(A)) used to determine reserves for group annuity contracts issued on the date as of which present value is being determined (without regard to any other subparagraph of section 807(d)(5)).” Recently (sometime in 2000, I think), at least 26 state insurance departments/commissioners adopted this table for group annuity reserving purposes. At that point, it became “prevailing”, and the IRS was required to specify it (or some variation) as the table to be used for this purpose.

There is no theoretical reason why the RP2000 table could not be used for current liability purposes. However, it is unlikely the insurance commissioners/departments of the various states will choose to adopt it. The RP2000 table was developed specifically for defined benefit plan purposes, while the GAR94 table was developed for the purpose of updating group annuity reserving processes.

An actuary will calculate life expectancy using a mortality table (chosen as appropriate for the purpose). In general, a mortality table consists of probabilities of death at each age. An actuary's training includes knowing how and why a table is constructed, to make sure the most appropriate table is used (for example, a table based on disabled persons would not be appropriate to use in valuing group annuity reserves). The actuary is also trained to use that table to calculate life expectancy, or joint life expectancy where a spouse might be included, or any other determination for actuarial value.

[imchipbrown]

PRP,

To see life expectancy in the GAR table, use an interest rate of 0%.

[flosfur]

PRP,

Might be presumptious on my part, but ..

Just in case it is not clear what Chip means by "use an interest rate of 0%", he means compute Annuity Purchase Rate for a given age at 0%. But be sure to use annual payment approach (not monthly payment) at the end of year 1, 2, 3, ....z (last age in the mortality table) - to my knowledge, this is the generally quoted life expectancy. If you want to be fancy (& more precise), add 0.5 to the number computed.

[Guest prp]

pax, Chip, flosfur,

Thanks for your responses, they were helpful and informative. Unfortunately, though, I don't have the resources to determine life expectancy using the methodologies described. What I am attempting to do is determine the lump sum equivalent of my DB plan annuity benefit at different ages. I have a calculator that will do that if I input the necessary variables (monthly or annual payment, monthly or annual interest rate, and number of monthly or annual payments - which equates to life expectancy). I have all the information that I need except I am unable to extract from the GAR 94 table the life expectancy of a person of a specific age. How can I determine from GAR 94 how many months or years a 58 year old, or a 59 year old or a person of any specific age is expected to live? Thanks!

prp

[david rigby]

We'll be glad to help. Need some information first. You do not have to tell us the amount of your benefit; we can give you the factor to use, then just multiply. (The following assumes you are covered in a pension plan that is not sponsored by a church or a governmental organization.)

1. we assume you are entitled to a pension benefit that is payable monthly. correct?

2. on what date (not age) is the benefit assumed to commence? (perhaps you have more than one date in mind.)

3. specify any early retirement adjustment that must be included (if not already included)

4. your date of birth.

5. the plan year (defined in the plan as the fiscal year on which plan records are kept); not necessarily the same as a calendar year.

6. the plan's definition of how it defines the interest rate for making such lump sum calculation (for example, "the rate in effect for the month prior to the beginning of the plan year") and the time period over which it remains stable (for example, the plan year or a plan quarter).

7. normal form of payment for your benefit, without adjustment, such as lifetime annuity, etc.

[Guest prp]

pax,

OK, item by item...

1) Yes

2) I'm considering 12/1/03 as a possible commencement date. However, depending on how interest rates move, it may be in my best interest to wait until sometime in '04. Let's assume 12/1/04 as an alternate date.

3) Early retirement adjustment is already comprehended.

4) 7/20/44

5) I don't know for sure, but I believe the plan year is the same as calendar year. See the answer to 6).

6) For a given calendar year (plan year?) the 30 year TSR as of August prior year is used. The 8/02 30 year TSR was 5.08% and that rate is used for all lump sum calculations for benefits commencing in '03. That being the case, any projections for '04 or later commencement dates will have to use an estimated interest rate. Consequently, I will be deciding on when to take the lump sum sometime after August of this year. I'll also be keeping my eye on the ongoing effort to replace the 30 year TSR as the basis for such calculations. (Which, by the way, would be an interesting topic for this board, somewhere. Decisions such as mine will certainly be affected by the outcome and timing of implementation. Are you aware if one yet exists?)

7) Lifetime Annuity

Thanks for your help and one additional ask. No one has yet been able to explain to me how you take the 94 GAR table and determing the life expectancy for a person of a specific age. Can you shed some light on this?

Thanks again.

prp

[MGB]

prp:

You have no need to determine life expectancy. The present value factor for determining lump sums does not make use of life expectancy. Instead, the calculation makes use of the probability of living to each age in the future, not just over life expectancy. Using life expectancy as the time period within a fixed time-period annuity calculator like you describe will only give you an approximation to the real present value you are looking for (and can be significantly off).

An "actuarial present value" (taking into account the probability of surviving to every year in the future) needs, at the minimum, a spreadsheet with one line of calculations for every year in the future.

However, using a present value approach with life expectancy as the input does have one advantage: It can quickly give you "relative" values with different interest rate inputs. For example, using that approach with two different interest rates and seeing the percentage change in present value that it produces is about the same percentage difference between the two correct actuarial present value calculations.

To determine life expectancy, given a mortality table in a column in a spreadsheet: Set up what is known as a "radix" (e.g., 100 or 1000, representing the number of lives a group starts with) at the top of the second column in line with the first age that you have a mortality rate. In the next line under the radix, set it equal to the prior line times (1 minus the mortality rate in the prior line). Copy that equation down the entire column. This column represents the number of people still living at each age, given that you started with a certain number. Then, sum that column starting with one line after the age you are concerned with. Divide that sum by the number of lives at the age you are concerned with. Due to this being at integral ages, adding .5 to the result will give you a better approximation to life expectancy. If I did this right (pretty quick without checking), life expectancy at 64 is 20.97 years.

Sticking that into a present value calculation at 5% per year, I get about 158 present value (times the monthly payment). However, using the actuarial present value approach in a spreadsheet, I get 145 present value, nearly a 10% difference.

More specific issues:

3) Most plans do not include the early retirement benefit in the lump sum. Most plans determine the lump sum as the present value of the deferred normal retirement benefit. You should check this out carefully as it can mean huge differences in the lump sum if there is a subsidy at early for the annuity benefit (which nearly all early retirement benefits are).

6) The change from the 30-year Treasury yield as the mandated minimum lump sum will not change for at least 4 or 5 years, and even then, will most likely be phased in over another bunch of years. There is currently only one proposal in Congress (Portman-Cardin III, HR 1776) to change it and it will not pass this year, and probably not next year either (election year). Adding all the phase-ins after that and I don't think we will have any real change for quite awhile.

[Guest prp]

MGB, pax,...

Thanks for the explanation of the actuarial present value methodology vs. the use of life expectancy in a present value calculation, it helps me understand, somewhat. That being the case, can you please figure what the present value factor ('x' times monthly payment) would be using 5.08% interest and an age of 59 years 4 months?

MGB, as to the specific issues you raise...

3) My plan calculates a full annuity benefit at age 62. If you commence benefits at an earlier age the full benefit is reduced by 5% for each year you retire early. I can make that adjustment to the monthly benefit to determine the amount to multiply by the factor discussed above.

By the way...are you aware of some present litigation on reducing benefits for early retirement in this (or similar) way?

6) Thanks for the input on the 30 year TSR issue..seems like I won't have to worry about it.

Regards,

prp

[MGB]

prp,

An immediate annuity factor at 5.08% at age 59 is 161.07 and at age 60 is 157.80. Your plan administrator may do various things with these factors. They may prorate between them for your specific age (very common) or they may use an "age nearest birthday" or "age last birthday" to use the factor at that age. Use of this factor assumes the plan uses method (1), below.

There should not be any litigation against an unreduced age 62 and 5% reduction before that. These are extremely valuable benefits that very few people have access to and would love to have such a subsidized early retirement benefit (i.e., much more valuable than the normal retirement benefit).

It is still not clear to me that you would get the present value of the early retirement benefit. The amount of the lump sum can be three possibilities:

1. The present value of the early retirement annuity multiplied by an immediate annuity present value factor.

2. The present value of the unreduced age 62 annuity multiplied by a deferred-to-62 annuity present value factor.

3. The present value of the annuity at Normal Retirement Age (defined under the plan) by a deferred-to-NRA annuity present value factor.

Even though you are eligible to take an annuity at an early retirement age, there is nothing in the law that requires the lump sum to recognize that early retirement benefit. Under most plans, when you elect the lump sum you are forfeiting the right to the extra value in the early retirement annuity.

You need to find out from the plan administrator which method they use to determine lump sums.

For example, using age 59 (and 5.08%) and a normal retirement benefit of 1000 per month:

Method 1 would be 1000*(1-.05*3)*161.07 = 136,910.

Method 2 would be 1000*127.82 = 127,820.

Method 3 (assuming NRA is age 65) would be 1000*99.78 = 99,780.

The 127.82 and 99.78 factors are for deferred annuities to age 62 and 65, respectively.

The fact that method 1 produces a larger lump sum than the other methods is why the early retirement benefit is called "subsidized". By using the early retirement benefit with the 5% reductions, the value is more than the deferred normal retirement benefit.

If the normal retirement age (not the unreduced retirement age) in the plan is age 65, there is nothing wrong with only providing a lump sum of 99,780, which results in you forfeiting the other 37k of value by not taking the early retirement annuity. This is how most plans operate.

[david rigby]

I agree with MGB's explanation and annuity values. If you need more decimal places:

immediate at 59: 161.07316

immediate at 60: 157.79831

immediate at 62: 151.05205

immediate at 65: 140.57747

deferred from 59 to 62: 127.81620

deferred from 60 to 62: 135.03101

deferred from 59 to 65: 99.78263

deferred from 60 to 65: 105.41503

BTW, not very likely that 5.08% will be valid for 2004 payments, probably closer to 4.75%. This could be significant: by postponing a possible retirement only 2 months (from 12/1/2003 to 2/1/2004), you might increase the lump sum by about 3%. But, that also depends on the factors outlined by MGB relating to rounding, how age is calculated, and possibly other administrative procedures.

[Guest prp]

MGB, pax,

Thanks much for your input. I'm now far more comfortable with the lump sum calculation methodology.

Regarding MGB's concern about how my plan would handle an early retirement calculation...

Based on the info that I get from plan documentation and some projections that I have received from the plan, I'm fairly certain that the number 1 methodology (The present value of the early retirement annuity multiplied by an immediate annuity present value factor) is used. When I did a 'back in' calculation with some numbers provided to me by the plan I arrived at a 'factor' of 159.7091 for a combination of age (59 yr, 4 mo, 12 days) and interest rate (5.08%). This seems to be a lot closer to methodology #1 than #2 or #3.

By the way, is there any straightforward way I would be able to refigure (or estimate fairly closely) a different immediate annuity present value factor by adjusting the ones you have provided (based on 5.08%) for a different interest rate and/or a slightly different age? I'm very much aware of the point made by pax that the 30 yr TSR in August may be lower than 5.08% in which case I will probably defer taking the lump sum until sometime in 2004 (depending on just how different the lump sum amount turns out to be). This will be the first thing I will be doing in early Sept. when the August rate is released.

In any event, thanks very much for your help.

Regards,

prp

[david rigby]

Only for purposes of a very rough estimate: to "convert" the 5.08% factors to those using 4.75%, multiply by 1.03.