February 17, 2012
By Nevin Adams, EBRI
In this business you are frequently asked “how much should people save for retirement?” Some try to answer that question with a degree of specificity that can be somewhat simplistic.
Let’s face it, even if those close enough to retirement to have a sense of what their pre-retirement income level is (and, flawed as that can be, most projections start from that assumption as a baseline for what you’ll want/need to spend in retirement—see “Replacement” Window), most struggle to turn that into a real savings figure.
Ultimately, of course, a reliable answer to that retirement savings question requires an understanding of the individual’s goals and/or financial needs—and, predicated on certain assumptions, there are any number of tools that can help individuals set a target and (based on that) establish a savings plan.
However, the planning question that almost never gets asked is: “And how certain do you want to be of achieving that target?”
Asked that question, I suspect most individuals would respond, “100%.” Unfortunately, much of the modeling that is being used to help individuals set those targets is based on averages: things such as average life expectancy, average investment experience, and—in the really in-depth models—average health care expenditures in retirement.(1) As a result, those models (useful as they might be in terms of framing a planning discussion) produce a result that will fall short…50% of the time.
In fairness, some of those shortfalls could be small. After all, if you’re a dollar short, you’re still short a dollar. But in some cases those shortfalls could be larger—much larger, in fact.(2)
And that’s a fact worth keeping in mind.
(1) For more information on these kinds of projections—and EBRI’s Retirement Readiness Rating—see the July 2010 EBRI Issue Brief, no. 344.
(2) For a more detailed discussion about those projected shortfalls—and how they can vary according to such factors as gender, marital status, and income levels—see the October 2010 EBRI Notes, Vol. 31, no. 10.