“Twenty to One”

The ratio of savings to annual retirement income is roughly 20 to one. Twenty-to-one is an easy ratio to remember. Promoting it would be cheaper and easier than requiring a falsely-precise number for retirement plan quarterly statements.

Advisors and individual investors often say they won’t recommend or buy income annuities today because interest rates and payout rates are so low. But would the average person know what constitutes a “high” or a “low” payout rate?

A recent finding from Cerulli Associates suggests that most people would not. In a survey earlier this year, Cerulli asked active 401(k) participants age 55 and older (“retirement income opportunity” clients) how large a lump sum they would be willing to pay for a lifetime income of $500 per month starting at age 65.

Assuming that the subjects were trying to come close to the actual cost of a $500/mo. annuity, rather than basing the estimate purely on their willingness to pay, then most of them wildly underestimated the necessary amount.

“Nearly three-quarters (72%) replied $25,000,” said a Cerulli report. Another 18% said $50,000. Only 3.8% came close to the right answer of $100,000. A handful (less than one percent) said $200,000.

“In reality, even among the most aggressive SPIAs, the same 65-year-old female would most likely need to spend between $90,000 and $100,000 to generate the $500, without receiving a death benefit guarantee,” Cerulli said. “This data also helps verify why annuities remain advisor-sold products and why less than 3% of 2012 VA sales were derived via the direct-to-consumer channel.”

Presumably, the Cerulli survey sample included many well-educated people. A high percentage of them should have been capable of deducing that $500 a month equals $6,000 per year. Had they thought about it for a moment or two, they could easily have seen that $25,000 would only buy about four years’ worth of income.

I don’t think lack of ability prevented them from finding a more plausible answer. They probably just weren’t used to doing that sort of calculation. In all likelihood, no one had ever taught them, challenged them, or forced them through the mental exercise of translating savings to retirement income. 

Given that less than half of adults (42%, according to the Employee Benefits Research Institute) have tried to figure out how much they need to save for retirement, the frequency of this type of error should come as no surprise. But it’s far from a harmless error. It will lead people to underestimate the cost of lifetime income. As a result, they’ll probably save too little, and perhaps run out of money before they die.

To help solve this problem, the Department of Labor has proposed that 401(k) plan sponsors and plan providers add a section to participants’ quarterly statements where this calculation has been done for them. Participants would see how much income in retirement (based on best estimates) their current savings (or possibly their projected savings at age 65, based on their current trajectory) would generate if it were used to purchase an immediate income annuity.

I’m not sure how effective this type of reform would be. Few participants look at their statements. Few people buy income annuities at retirement. Retirement plan sponsors have already tried to show participants that a slight increase in their contribution rate can produce a significant increase in their savings over time. I haven’t seen any evidence that these efforts have been wildly successful.

Maybe there’s a cheaper, easier way to educate people. There’s already a rule of thumb that says a retiree can spend roughly 4% to 6% of savings per year during retirement, depending on the markets and whether they use a systematic withdrawal method or buy a life-contingent annuity.

We can therefore say that the ratio of final savings to annual income, at current interest rates, is roughly 20 to one. Twenty-to-one is an easy ratio to remember; why not publicize it? Put it on billboards along the nation’s highways. Paint it on the sides of barns in farm country. If more people knew and used that heuristic, a much higher percentage might have been able to answer Cerulli’s survey question more or less accurately.  

© 2013 RIJ Publishing LLC. All rights reserved.