Your Retirement: Bet in a Guessing Game?

For years, financial advisers have given their clients authoritative, seemingly precise guidance to tell them exactly how much they need to save for retirement—and how much they can spend when they get there. Last year’s market crash has exposed some flawed assumptions behind those numbers.

These complex formulas or “models” have dominated retirement-planning for a generation; advisers enthusiastically adopted these tools, and millions relied on them. But now that the steep decline in stocks has blown a hole through many retirement portfolios, even the experts who design these models are acknowledging that they have serious limitations. Indeed, some insiders say the number crunching is just an educated guess. “A really sophisticated guess, but a guess nonetheless,” says Gregg Janes, a developer at EISI, a company that makes financial-planning software.

Of course, financial planners aren’t fortune-tellers. Few could have predicted the speed and depth of the downturn. For many, their withdrawal calculations are simply broad guidelines for clients. But critics say that the numbers create a facade of authority that’s propped up by some relatively fuzzy math. Even planners who use the figures say they don’t do enough to account for the impact of something like the crash. That’s one reason national brokerage Edward Jones doesn’t give its advisers statistical models at all. With the most commonly used models, it’s too easy to misinterpret the results, says Scott Thoma, the firm’s director of investment advice: “We think it instills false confidence.” And even defenders of the models are rethinking them; according to a recent survey by the Financial Planners Association, almost 50 percent of advisers say they’re reconsidering how much they’ll tell clients to spend next year. The models are robust tools, said Matt Sommer, director of financial planning at Smith Barney, “but in 2008, they just didn’t work out.”

With markets having tumbled so far from their peaks, investors and their advisers are struggling to figure out how their portfolios can recover. Often, they’re using the same models they relied on in the first place—essentially, they’re guessing again. But elsewhere, planners, economists and computer whizzes are debating how to refine, improve or even throw out their formulas.

If you’ve made the journey to an adviser or broker for retirement-planning help, you know the drill. The adviser translates the client’s retirement fantasy into a cash target—how much money it will take to cover expenses each year. For younger investors, the adviser uses the figure to calculate how much the client needs to save. For those already retired, it’s the source of the withdrawal rate.

Three years ago, Ken Miller, owner of a Cleveland-area auctioneering firm, crunched those numbers with his adviser, tallying his income and assets—Social Security and a $2.5 million portfolio—and balancing them against his housing and medical costs and his desire to splurge on the kids and grandkids. Click-click on the keyboard and voilà: Ken and his wife, Bonnie, could ride their nest egg through retirement on about $125,000 a year in income.

Ken, feeling confident, stopped taking a paycheck and started to relax. Then came the crash, and Ken’s investments lost about 50 percent just as he was turning 65. Now his adviser says he has to cut his expenses by about $40,000 this year. That means no shopping sprees, no cruises and fewer dinners out. On the other hand, if the Millers make these cuts now, they should have more to live on in the future. “I certainly understand the math,” Ken says.

Still, that’s the kind of math some advisers are starting to question. A planner would typically run a computer model that tests any withdrawal plan under hundreds of different market conditions. The result is expressed as a probability: Stick to the plan and there’s a 90 percent chance you’ll have money left over in 30 years. Put another way, nine times out of 10, you should be fine.

So far, so uncontroversial. But in practice, the process requires a slew of assumptions about the markets, the economy and the client’s behavior—and this is where the problems begin. For starters, there’s a lot of incentive to be optimistic. If a client agrees in theory to spend less, he can retire sooner; if he and the planner decide that the markets will perform well, the client can save less or spend more. In UBS’s planning-software framework, to take one very typical example, it’s assumed that a “moderately aggressive” portfolio will return 7.5 percent per year; if the planner takes only a slightly more pessimistic view of the markets, the chances of the client running out of money can rise from 10 percent to more than 25 percent. UBS says it’s up to the planner to vary the assumptions “to help clients understand a range of possible outcomes.”

The software itself raises its own serious questions. The most popular model today is “Monte Carlo” analysis (named after, yes, that Monte Carlo). It’s a powerful statistical tool; the oil and gas industry, for example, relies on it to project future supply. But corporate America’s supercomputers can run millions of simulations; in contrast, some advisers’ desktop software uses as few as 150. The fewer the trials, the less likely the software will account for unlikely events like the 2008 crash, acknowledges EISI Senior Vice President Linda Strachan. And even if a planner runs millions of trials, Strachan says, a flawed assumption, like an inflation projection that’s too low, “will give you a detailed examination that’s completely wrong.”

These pitfalls turn the models into an increasingly cloudy crystal ball. Scot Stark, a certified financial planner in Freeland, Md., whose clients together have about $10 million in investable assets, has a Monte Carlo system that runs thousands of simulations, but he compares using it to “driving forward looking in the rearview mirror.” Nonetheless, he hasn’t found anything better for figuring out an income plan. So he keeps using it, while reminding clients that “there are no guarantees.” He doesn’t get into explaining its limitations unless a client asks him directly, and he can’t remember a time when this has happened.

Clients may not question the numbers, but plenty of critics do—and they say that the recent market bloodbath only proves the model’s shortcomings. Perhaps worse than any software issue is the false certainty implied by the hard numbers of a Monte Carlo–style model. “It says I have a 98 percent chance of success—I must be home free,” says Edward Jones’s Thoma. “But that’s not what it means at all.” In their defense, brokers and firms say their models aren’t perfect, but they’re the best they’ve got. Smith Barney’s Sommer says the Monte Carlo analysis itself isn’t broken: If you have a 90 percent chance of success, and this is the other 10 percent, he says, “it’s not the Monte Carlo’s fault, it’s not the adviser’s fault, it’s just bad luck.”

As they try to refine the system, some advisers are putting clients on a drastic spending diet. A year ago, JPMorgan Chase’s proprietary models gave its advisers the go-ahead to let clients withdraw 5 percent of their portfolio each year; in the past, they’ve okayed spending as high as 10 percent. Today, the firm typically uses 3 percent as a starting point. Since the collapsing markets have shrunk most clients’ portfolios, such a move could result in a retirement pay cut of more than 50 percent. That’s obviously unpleasant news, says Mary Ann Sisco, national director of wealth advisory for the firm’s private wealth management division. But she estimates that a retiree who stayed at the 5 percent rate would have only a 30 percent chance of avoiding going broke. Thank goodness for the models, she adds. “We can show them the numbers and explain it.”

Others are relying on smaller tweaks. Retirement-plan giant T. Rowe Price follows one of the most traditional guidelines, the 4 percent rule. The idea: In year one of retirement, you withdraw 4 percent of your portfolio; each year after that, you increase the dollar amount slightly for inflation. Four percent is seen as the “hell or high water” rate that bad markets can’t bust. But Stuart Ritter, a T. Rowe Price financial planner, calculates that due to the crash, the success rate has dropped into the low 70 percent range for someone who retired in 2008. He suggests giving up the inflation bump for three to five years. That’s a relatively small sacrifice that doesn’t mean big lifestyle changes—as long as there isn’t another shock on the horizon, like more investment losses or an unforeseen expensive emergency. Ritter acknowledges as much. “Life is unpredictable,” he says. “But it doesn’t mean you can’t start out with a good, solid plan.”

Some folks are thinking outside the piggy bank—moving away from the idea that retirees can milk predictable income from a stock-and-bond portfolio. Ed O’Connor, managing director of wealth management at UBS, is among those encouraging clients to rely on a combination of Social Security, immediate-income annuities and (for the lucky few) pensions to cover essential expenses like housing and medical care. That frees them up to be more aggressive with other investments—so that, in good years, a retiree could keep up with inflation while still affording the occasional splurge.

Jonathan Guyton, a Minnesota certified financial planner and researcher on spending in retirement, takes the flexibility argument even further. He says that a retiree can spend 6 percent or more of his portfolio in a year when markets are down—essentially assuming that it’s only a matter of time before things improve. After all, history tells us that the market will rebound. Markets “would have to be dismal for eight to 10 years for [that strategy] to blow up,” he says. “And I don’t think we’re there.”