We put five programs through their paces. How much should you rely on them? By David Landis, Contributing Editor July 31, 2007 Bill Rudy's financial life isn't especially complicated. The 43-year-old project manager is single, has no kids and sold his house last year. He expects his employer's 401(k) plan (he works at a bank) and an Air Force pension to cover his retirement needs. What he wants to know is how soon he will have saved enough to begin his retirement. He has tried to get an answer from a half-dozen online retirement-planning tools, but they have hardly clarified the picture. He now has competing estimates of his savings at retirement that range anywhere from $500,000 to $2 million. Each tool asked for different pieces of information, and each produced a different answer. "You'd think these guys would all use the same basic financial computations," muses Rudy, who lives in Richmond, Va. "How or why they come up with such different answers, I don't know." RELATED LINKS Five Tools Put to the Test How to Find a Financial Planner We had the same experience. We created a simple financial scenario in which a young couple -- call them Jim and Jenny -- have amassed $100,000 in Jim's 401(k) retirement plan, to which he will continue to contribute $12,000 a year over 30 years (he also receives a $1,000 employer match). We ran our criteria through five online planning tools (though not all were the same ones Rudy used), and we wound up with five radically different projections of the total savings Jim and Jenny would have at retirement and their income afterward. Which tool gave the most accurate results? We have no idea. Only time will tell whether these 30-year forecasts have any validity. What we can say, after quizzing some of the providers of the tools and digging into the fine print of how they work, is that retirement-planning software is as much art as it is science. Although some programs have impressive graphics and can calculate thousands of scenarios in just seconds, we didn't come away with a high degree of confidence in any of their findings. Advertisement About probabilities No piece of software can predict the future. The best it can do is calculate probabilities based on a long list of assumptions. These assumptions can include such things as the rate of inflation, investment returns, tax rates, spending needs and how long the user will live. Fidelity's Retirement Quick Check tool, to take one example, requires 28 dense pages to explain how it chooses and weighs each factor; small variations can lead to widely disparate results. For example, adding a single percentage point to an assumed 6% return on $100,000 can mean a difference of nearly $187,000 in your forecasted nest egg over 30 years. "It's like figuring the course of the space shuttle on the back of an envelope," says Boston University economics professor Laurence Kotlikoff, designer of one of the tools we tested and a critic of most of the others. "If you're off a few degrees, you'll end up on Jupiter." Most of the tools we tried allowed the user to adjust some key assumptions, such as future interest rates, investment returns and life span. However, most investors are hardly qualified to forecast economic factors over the next six months, let alone decades into the future, and none of us knows how long we'll live. On these matters, you're better off accepting the tools' built-in assumptions (generally, 3% inflation and 6% investment return), and planning for a very long life (of, say, 100 years). A more controversial assumption concerns the amount of money you're likely to spend in retirement. Quick Check assumes you'll need 85% of your preretirement income after you retire. Financial Engines assumes 70%. Both presuppose that a retiree's spending increases at the rate of inflation throughout retirement. But that's a point of contention. Ty Bernicke, a financial planner in Eau Claire, Wis., says studies show that retirees actually spend less than they did during their working days and that their spending decreases for everything, except health care, the longer they are retired. "Retirees don't take a lot of vacations -- they don't have the energy they used to," says Bernicke. "They don't eat as much, and they don't drive as much." He believes retirement-planning tools could encourage people to save more than is necessary. Advertisement Marketing gimmicks? Some critics say financial-services companies use their retirement-planning tools as marketing devices and deliberately set the so-called income-replacement ratio far too high in order to encourage people to invest more money with them. Not surprisingly, James Cornell, a senior vice-president at Fidelity, disagrees. He argues that the demise of traditional pension plans, doubts about the future of Social Security and soaring health-care costs have combined to dramatically decrease retirees' sense of financial security. "I don't think we're trying to get people to save more," he says. "We're trying to get people to be informed, to have a plan." Another complicating factor is the methodology some tools employ to estimate future investment returns. Rather than use a simple mathematical formula to calculate the impact of, say, an 8% return and 3% inflation over 30 years, the more advanced tools use what is called a Monte Carlo simulation. This is a computer simulation that can quickly run thousands of "what-if" scenarios. So, rather than assume an unchanging 8% annual return over 30 years, a Monte Carlo simulation can factor in a range of possible returns as well as a range of inflation rates and other factors. The results are expressed in percentage terms. If a portfolio achieved, say, an 8% return in 900 out of 1,000 scenarios, a user is said to have a 90% chance of achieving that result. Different tools, however, build different ranges of assumptions into their Monte Carlo simulators. Plus, they may build in varying projections of Social Security income. In addition, although Fidelity, Financial Engines, Morningstar and Kotlikoff's ESPlanner allow you to enter the details of your actual portfolio, their projections are based, for the most part, on average projected returns for stocks, bonds and cash. The tools come with different methodologies for determining how long a portfolio will last in retirement. Morningstar's Asset Allocator, for example, assumes retirees will be extremely conservative, so it builds in a gloomy 1%-a-year after-inflation return on investments after retirement. Advertisement Finally, every tool presents its findings in a different way, making it difficult, if not impossible, to compare the results. Financial Engines and Morningstar give their results in current dollars, which aren't adjusted for future increases in prices. Fidelity and Choosetosave.org present findings in future dollars, which do take into account the effects of inflation. ESPlanner, the program developed by Kotlikoff, the Boston University professor, uses a methodology that is radically different from that of the other tools. Rather than assuming that your savings peak at retirement, after which you spend them down, his program calculates a constant level of spending, both before and after retirement, that will accommodate your goals. Kotlikoff says this approach assumes that people prefer to maintain a constant standard of living rather than make big changes according to how much they've saved at retirement. Shock therapy So how should you use these programs? If you have neglected to save for retirement, running the numbers through these tools can serve as a kind of shock therapy to persuade you to get serious about saving. "I don't have any problems with using these programs as a starting point," says Lisa Bolton, an assistant professor of marketing at the Wharton School. "If you're starting from zero, and you know nothing, they're better than nothing." If you have a systematic retirement-savings program already in place, there's certainly no harm in using the results as a ballpark estimate of your progress. Tools from Fidelity, Financial Engines and Morningstar allow users to adjust their portfolio holdings to see what effect the changes could have on future returns. The tools can also suggest ways to tweak your portfolio to improve its performance. Advertisement In fact, Financial Engines, which is available to many participants in 401(k) plans, offers some of those employees the opportunity to tweak their portfolios using the tool, then apply the proposed changes to their actual retirement portfolio with the click of a button. However, you should be wary of putting that much faith in a technology that is clearly in an early stage of development. Turning to live help In fact, we think Bill Rudy, the Richmond project manager, has the right idea. He's thinking of hiring a professional financial planner to help him develop a retirement plan. Ironically, it's possible that the planner will use his own software program to help plot Rudy's financial course. If so, the planner should be familiar with the assumptions and biases built into the tool, and know how to interpret its findings. Furthermore, the planner should supplement the findings with his or her own real-world experience. As for the rest of us, until these tools have been further refined, the most prudent strategy is to save as much as we possibly can for retirement. Even Bernicke, the Wisconsin planner who is highly critical of these tools, allows, "I don't think many people look back and say, 'Boy, I wish I would have saved less.'"