Friday, April 21, 2006

Pretzel Logic

Awhile back, after receiving the first part of the "actuarial reduction tables" Mercer developed for PERS, I questioned various PERS senior officials about them. In the main, my query involved the "life expectancies" implied by the tables. In my reading of the tables, it appeared that Mercer was assuming extraordinarily long lives and, that by so assuming, PERS would never collect the full amount. I was quite skeptical that this was the intent, and raised a hypothetical with PERS. In one response, PERS told me that "...when you factor the impact of COLAs on the reduced benefit, it brings the payback period in line with current longevity assumptions". I responded: "...how do COLAs on the revised benefit factor into anything related to the payback since the payback is based on amounts allegedly paid illegally, frozen on a date certain, and to be repaid based on actuarial life expectancies." I further wrote: "...if Joe Blow owes PERS $5000...is 60 years old, and .... will live to 80 years old, then Joe would pay back $5000/240 [$20.83] per month no matter how long he lives." And finally I concluded: "...the revised tables suggest that Joe is going to live to 98, virtually guaranteeing that he will never repay the full amount (unless the $5000 which is owed is actually inflated by a 2% COLA that he isn't receiving and, according to PERS, isn't entitled to receive.)"

I've waited patiently for PERS to get back to me on this. Earlier this week, I received an email from PERS which contained "the" answer to my question, direct from Mercer. I'm going to quote it directly in italics to set it off from the rest of the text: "The overpayment is a fixed amount that is collected over the expected lifetime of the retiree and his or her beneficiary with no interest charge. The method of collection is to reduce the benefit the retiree would otherwise be entitled to. Because this benefit amount is entitled to an annual COLA, by reducing it, we are also reducing future COLAs. By including an assumption of a 2% COLA in the development of the reduction factor, we are taking into account the reduced future COLAs as a part of the repayment so that we do not collect too much from the retiree. So, the retiree is not making level payments on the overpayment, but is making payments that are assumed to increase 2% per year. Using your example of a retiree who owes $5000 and is expected to live 20 more years, monthly payments would start at about $17 per month, but would be expected to increase 2% each year. The starting payment is less than the method you proposed ($5000 / 240 = $20.83), but with the 2% increases becomes about the same after 10 years and collects the same $5000 after 20 years."

It took me quite a bit of time to parse carefully what Mercer is saying, but I finally think I understand it. The actuarial repayment tables are an artifact of the 2% COLA. What Mercer has proposed is that we will be paying back a fixed amount each month, determined by the actuarial factor at the time repayment begins. Although we will not "see" this, the repayment amount is "assumed" to increase by 2% annually, although it will not, in fact, increase. By "assuming" a 2% annual COLA on the payment, it is also assumed that by the time we reach our TRUE actuarial life expectancy, the original amount we owed would be paid back. I suppose one could view this as good news because the actuarial reduction is smaller than common sense and simple math suggest, but it actually changes nothing significant. If you die early, you pay back less; if you reach an older age than actuarially expected, you pay back more. The only consolation is that you pay back in dollars that are deflating by the true cost-of-living less the 2% COLA assumption, and your monthly amount is fixed at a lower amount from the beginning.

Addendum: If you are concerned about the implications of this explanation, you should definitely express them to PERS, to its Board, and to the PERS Coalition. I plan to make a portion of the Actuarial Repayment Factors available shortly. I have only those for Option 1, 2, and 3 retirements. I don't have those for Option 0, 2A, 3A, or 4. You can use my "Lipscomb Calculator" to get some rough idea of how much PERS thinks you'll owe (use this as a guide, not as a statement of fact), and what your adjusted benefit will be (without actuarial reduction). Then use the actuarial reduction factor table for your retirement option to determine the "factor". That will tell you what your monthly repayment amount will be. The hard part (for some) is to take that starting amount and increment it by 2% annually for as long as YOU expect to live based on your known history. That will tell you how much you're actually repaying over your lifetime.

No comments: