Retirement Models
That Let Reality Bite

What happened to stocks during the past three years?

a) They soared in 1998 and 1999 and slumped in 2000.

b) They posted a three-year average annualized gain of 12.3%.

Both are true. The first answer, however, captures our sense of the market's craziness, while the 12.3% average seems willfully misleading. Yet, when folks calculate how much money they might have in retirement, they often ignore the craziness and rely on long-run averages.

That could turn out to be a big mistake. "People don't realize that averages can be meaningless," says Moshe Milevsky, a finance professor at York University in Toronto. "It's how you get there that counts."

Even now, most retirement-planning calculators are what experts call "linear" or "deterministic." They assume folks will earn the same return year after year. But in truth, the world is a lot messier than that, with returns coming in an unpredictable mix of dazzling gains and rotten results.

To capture this uncertainty, Wall Street is turning to "probabilistic" models, which can help investors gauge their strategy's chances of success. (Many of these models use so-called Monte Carlo analysis. But despite the hype about Monte Carlo, you don't always need it to calculate probabilities.)

For instance, T. Rowe Price Associates' Web site (www.troweprice.com) offers a retirement-income calculator, while Financial Engines (www.financialengines.com) and Morningstar's ClearFuture (www.clearfuture.com) can help those saving for retirement. These models won't offer a simple thumbs up or down on your strategy. Instead, they focus on how likely you are to meet your goals.

Intrigued? Here is why you might want to try playing with these new probabilistic models:

Deterministic models give investors a false sense of security, argues Scott Trease, a financial planner with Ronald Blue & Co. in Charlotte, N.C.

"When I do linear analysis using average returns, I get a single number for ending wealth," Mr. Trease says. "But when I run a Monte Carlo simulation, I might find that I only achieve that ending wealth 45% of the time."

Probabilistic models can help you fine-tune your retirement saving-and-spending strategy, so that you have a greater chance of success. But don't expect to eliminate all risk. For your strategy to have a 100% chance of success, you may have to save extraordinary amounts or spend pathetically little.

"As a rule of thumb, I like a 90% success rate," Mr. Milevsky says. Beyond that, "it takes a lot of effort to reduce risk further."

Deterministic models not only overstate your chances of success, but also don't show how badly you could fare.

According to Chicago's Ibbotson Associates, stocks have gained 11% a year during the past 75 years. But even if you enjoy average returns that approach that level, you may end up with far less wealth than a deterministic model suggests.

As you discover when using a probabilistic model, it is entirely possible that long-run returns could be one or two percentage points lower, which would make a big difference to how much wealth you accumulate. More critically, even if you clocked 11% a year, a lot depends on when those gains are earned.

If you are saving for retirement, you would rather have lousy performance now and great returns later, when you have more money at stake. Conversely, if you have just retired, you want great returns now, before your portfolio is depleted by your own spending.

"The sequence of returns is critical when it comes to spending in retirement," says John Rekenthaler, president of online advice for Chicago's Morningstar Inc. "If you get bad returns early in retirement, you can quickly end up with just 75% of your starting assets."

If you plug your current wealth, monthly savings, time horizon and expected return into a deterministic model, you will be told exactly how much you will amass. Not enough money? The temptation is to gun for higher returns, by plunking more into stocks.

But there are two problems with that. First, you may not be able to stomach the extra risk. Second, that extra risk may go unrewarded, notes Alexander Zaharoff, a managing director with J.P. Morgan Chase & Co.'s private bank in New York.

"In a deterministic model, taking more risk by putting more in stocks will always give you a better outcome," he says. "But with a probabilistic approach, taking more risk by taking on more stocks may reduce the chances of staying solvent. It may raise the upside, but also increase the probability of running out of money."

To understand this trade-off, consider a rule of thumb offered by William Bernstein, an investment adviser in North Bend, Ore. Mr. Bernstein notes that stocks have a higher expected return than bonds.

But there is a cost. He figures that, over the long haul, bonds could potentially fall one percentage point below their expected return, while stocks might fall three percentage points below. The bottom line: Adding more stocks may boost your portfolio's likely performance. But because there is more uncertainty about stock returns, there is also a greater chance of earning far less than you expected.

Write to Jonathan Clements at jonathan.clements@wsj.com