The following is a continuation of a discussion from last week, which was in response to Joe Tomlinson’s article, We Need a Bold Solution to Fix the Retirement System, which appeared on October 9:
Dear Editor,
I would like to raise some additional questions with regard to Joe’s analysis. Specifically, I assumed that less has to be saved (on a per-person basis) to comfortably provide for retirement income for a group of 10,000 individuals than for a single individual. For the group we can assume a distribution of life expectancies and hence payouts with a mean life expectancy (say 80 to 85 for a group of 65-year olds) that is significantly less than any single individual must assume for themselves (say 95 to 100 if no health/family history/exposure reason to assume less). This should translate into a lower required savings rate for the group to cover retirement costs than for the individual.
- Do my assumptions make sense? How much less would have to be saved on a group (versus individual) basis?
- I assumed there cost of inflation-adjusted annuities could serve as a basis for calculating the savings differences. I assumed that if a competitive bid was sought for an annuity for everybody in the group (e.g., all employees retiring from a large company), it would be significantly cheaper than the cost of an individual annuity wherein the insurance company by default must assume adverse selection (i.e., that the purchaser will live significantly longer than average).
- If this annuity analogy is inappropriate, than maybe a separate calculation is required to estimate the difference in savings required for a group (with average life expectancy) payout versus the per-person savings required for an individual. Would a Monte Carlo analysis of these respective savings rates with the same confidence intervals illustrate the difference?
For this reason, a system that provided a government, business or association a role in consolidating retirement savings for groups of individuals (whose participation in the groups would be mandatory) would require less savings than the current individual-based 401(k) system.
Let me raise a related issue that argues for a group if not governmental approach to retirement savings, as opposed to the individual approach embodied in 401(k)s. When I meet with clients, I point out to them that they might do everything “right” with respect to their savings and investing and still fail because they were born at the wrong time or retired at the wrong time (e.g., compare an individual who retired in 1972 to one who retired in 1982). The variability in market returns, interest rates and inflation rates forces the individual (who only has one life and one shot at success) to significantly over-save. Thus I tell my clients that my goal is to level out their happiness in life so, when they are young, they do not over-save (resulting in money they cannot enjoy when they are 85) or under-save (resulting in them eating cat food when they are 85). Yet, for practical reasons, I tend to push to the over-saving side.
The result is that those who ‘’appropriately” save will on average end up with a greater legacy than intended. This issue can only be addressed in the context of group savings (as discussed above but with the group encompassing individuals retiring over a large range of years – longer than major economic cycles) and involvement of an outside entity – probably the government, but possibly a non-profit insurance company or a group-pension entity – that had the charge, time horizon and resources to try to average out those long-term economic variables. The purchase of mandatory group inflation-adjusted annuities does not address this issue of variation in economic climate when the group annuity is issued.
All the best,
Barry Korb, CFP®
Lighthouse Financial Planning, LLC
Potomac, MD
Joe Tomlinson replies:
The short answer is that, in the article, I made the assumption that, at retirement, DC balances would be converted to inflation-indexed immediate annuities. So I've already included the pooling advantage of annuities that Barry points out. If individuals didn't annuitize they would need even larger balances than I showed in order to be safe.
Barry raises the question of whether I was too conservative in my estimate that, if we relied on DC plans only, contributions at roughly the 15% level would be required to insure a safe retirement. More specifically, he asks if I considered the benefits of annuitization and whether group annuities might bring even further benefits.
In my analysis I did include the benefit of annuitization because I assumed that DC savings balances at retirement would be used to purchase individual single-life annuities. I assumed a payout rate for an inflation-adjusted SPIA of 5.20% (reflecting long-term average interest rates rather than today's low rates that would produce payouts of around 4%). If I had not included annuities, I would have assumed a safe withdrawal rate of around 3.5%, which would have multiplied my 15% estimate by 5.2/3.5 and produced required contributions of over 20%. So there is an advantage to annuitization, but I already included it.
Barry makes the additional point that group annuitization would produce more favorable outcomes than individual annuities. Indeed, I did use estimated individual annuity rates. My estimate is that if we required everyone to annuitize on a single-life basis (and that would include even individuals with severe health problems), that change would bring my 15% down to about 13% – still way higher than today's average savings rate.
But I emphasized "single-life basis." I haven't adjusted my numbers to reflect that a high proportion of annuities would likely be sold to couples and that such rates (individual or group) would need to be lower to reflect joint-life expectancy (A 5.2% single-life rate would drop to close to 4% on a joint-life basis). My estimate is that doing the further refinement to include a high percentage of couples would more than offset this group advantage, so the overall number would come out above 15%.
My basic message is that, if we rely on only DC plans, the required savings rates for individuals to assure a safe retirement are well above what participants are saving today (9.6% on average for those who participate, and only 42% participate). This conclusion holds however we refine the numbers.
