Skip to content
Link copied to clipboard

Retirement calculators don't factor in risk of outliving assets

Retirement calculators have many flaws, and now a new study finds yet another glitch in the ubiquitous tools that help Americans figure out how much to save for retirement.

BOSTON - Retirement calculators have many flaws, and now a new study finds yet another glitch in the ubiquitous tools that help Americans figure out how much to save for retirement.

Apparently, too few calculators take into account longevity risk - the risk of outliving your assets - when they crunch the numbers.

Most calculators "are limited in the way they handle longevity risk," wrote John Turner, the study's author and director of the Pension Policy Center.

Turner identified 25 Internet-based retirement calculators and then took those programs through their paces using two fictitious profiles, one designed to tilt the results in favor of annuitization or partial annuitization and the second designed to tilt the results against buying an annuity.

He found that most programs are designed to figure out whether the user has enough money and retirement income to provide a desired level of income for a fixed number of years. The programs do not, however, consider longevity risk and the value of annuitization, he wrote in his paper, published by the Pension Research Council.

Some programs cope with longevity risk by setting the fixed planning period as late as age 95 or 100. But that's not good enough, especially if you have a long life expectancy. In those cases, "an annuity might be a more efficient way of dealing with life-expectancy risk," wrote Turner, who has studied retirement software and annuitization extensively.

Also, just one of the 25 programs suggested annuitization would be sensible for the first profile. Turner said nearly all of the programs studied suggested phased withdrawals as a way to cover a fixed lifespan. Two programs mentioned annuitization, but more as a general response, not specific to the data entered. And three programs suggested annuitization, but only in supplementary materials.

In short, the programs, some of which can be found on insurer's websites, don't advise annuitization - and that may be among the reasons why people do not annuitize.

"We believe that one of the reasons why few people annuitize is that they are generally not advised to do so in online retirement planning calculators," Turner wrote.

Should retirement calculators be blamed for the low use of annuities? Should you avoid using these online calculators altogether? Should you use them knowing full well that the recommendations might be incomplete? And should makers of such calculators start incorporating longevity risk into the code?

Don't blame calculators for low annuity use

It's not quite fair to blame retirement calculators for the lack of annuitization in retirement, said Anna Rappaport, president of Anna Rappaport Consulting.

"We do not know how many people base their decisions on recommendations by retirement planning software," she said.

Others agree. "There are many reasons why people don't annuitize," said Cathy Weatherford, president and chief executive of the Insured Retirement Institute.

According to Rappaport, many Americans and advisers simply don't like or know enough about annuities. She said sponsors of defined-contribution plans that offer annuities say that the take-up rates of annuity options are extremely low. And in defined-benefit plans that offer lump-sum payouts, those are usually chosen. Plus, advisers in some circles tend not to recommend annuities very often either, she said.

To be fair, Turner said in a telephone interview that retirement calculators are not the only reason why there's so little annuitization in the U.S. Rather, it's just one of many reasons, he said.
Avoid using calculators?

It would be easy to decide to avoid using Internet-based retirement calculators altogether after reading Turner's study and his previous work on the subject. Such calculators don't factor longevity risk or Social Security adequately into the mix. What's more, these programs tend to suggest phased withdrawals, which, Tuner said, exposes people needlessly to longevity risk. And the list goes on.

Still, Turner said it's better to use these calculators, despite their flaws, than not use anything at all. "It's a complicated problem and some information is better than none," he said.

Others agree. "If used for informational purposes or self-education, I think these calculators can be useful," said Eric Harrell, of PIEtech, the makers of MoneyGuidePro software. "I do not, however, agree that solely basing a decision on the information provided from them is always a good idea."

Harrell said users of Internet-based retirement calculators should understand the calculator's underlying assumptions and how the calculator is implementing the data being entered. "This is in addition to understanding the result," he said.

Turner said the purpose of his study was less about being critical of Internet-based retirement calculators and more about seeking ways to make them better.

One way to improve such calculators would be to factor longevity risk into the equation, Weatherford said. "It's a huge risk for baby boomers," she said.

That might be easier said than done. "Depending on the results produced, Monte Carlo, straight line, or deterministically examining varying returns, the implementation of longevity testing may provide skewed results," Harrell said. "What is a sufficient longevity test? How long is too long to examine? How does varying the planning age affect the results?"

But even if longevity risk isn't factored in to Internet-based retirement tools soon, Turner said he'd be happy with this improvement alone: "A more complete approach would be to present the annuity benefit that could have been received, given a person's retirement age and accumulated assets. Alternatively, a program could present both a phased withdrawal solution and a partial annuitization solution, and compare the two."

(c) 2010, MarketWatch.com Inc.

Distributed by McClatchy-Tribune Information Services