Science Watch

New enrollees in company 401(k) plans must choose from a dizzying array of options. Should they invest in stocks, bonds or property investments? How much should they contribute to each?

In a perfect world, investors would answer these questions by taking stock of their risk tolerance and savings goals, says Henry Kaelber, the president of a Virginia-based investment management firm. However, a new study, published in December's Journal of Experimental Psychology: Applied (Vol. 12, No. 4), finds that people are swayed by irrelevant factors. For instance, investors seem to be influenced by the proportion of options offered-putting more of their money in stocks, for instance, when more stock funds are available, notes study author Jörg Rieskamp, PhD, a researcher at Germany's Max Planck Institute for Human Development.

What's more, most people don't learn from their investing experience, found Rieskamp. The pros know to look back at the historical performance of a fund, but the average investor seems to respond only to the most recent returns, he says.

"People think they can observe trends or patterns and predict the future returns...but this is hard to do," Rieskamp says.

Kaelber agrees, adding that even confident investors often make mistakes.

"People who characterize themselves as conservative more often than not have a far more risky grouping of investments than they realize," Kaelber notes.

Simulated experience

Rieskamp explored the question of how people determine their 401(k) allocations with 80 students at Indiana University Bloomington. With an average age of 22, the students are a good population to study since they are the 401(k) investors of the near future, Rieskamp says.

The researchers gave the participants $5,000 in play money to invest in four securities, as part of a retirement savings plan. Half the participants had two stock funds and two bond funds to choose from, while the other half had three stock funds and one bond fund as investing options. Participants also reviewed information on the average return and variability of the funds.

Many of the students distributed their contributions about evenly across their four different options: Those who had three stock funds and one bond fund to choose from put 75 percent of their savings in stocks and 25 percent in bonds, on average. The students who had two stock funds and two bond funds on the menu generally put 50 percent of their money into stock and 50 percent into bonds.

After the investing task, Rieskamp measured the students' risk tolerance by giving them a series of choices where they could either receive a small amount of play money, or the chance to gamble for a larger amount. After the experiment, participants received 10 percent of the money they "won" during the risk-measurement task and a percentage of their retirement savings.

The students with more of a penchant for gambling did tend to invest more in stocks than bonds. However, "this effect was not that strong in comparison with the composition effect," Rieskamp notes. Even risk-averse participants tended to invest heavily in stocks when many stock funds were offered, he says.

Real money

In real 401(k) plans, people often have dozens of funds to choose from, notes Gur Huberman, PhD, a finance and economics professor at Columbia Business School. And contrary to Rieskamp's findings, a recent study of 572,157 401(k) participants, conducted by Huberman and his colleague Wei Jiang, PhD, found that people's fund allocation is not affected by the proportion of different kinds of securities offered by the plans. That is, people with a 401(k) provider that offers many stock options-in the real world-aren't more likely to put more money in stocks, according to the study, published in the April 2006 issue of the The Journal of Finance (Vol. 61, No. 2).

Because the fund selection wasn't affected by fund composition, there's no reason to assume that people weren't making smart choices, says Huberman.

"They are investing in diversified funds, which strikes me as pretty reasonable," he says.

However, after people chose their securities funds-three to four on average-they tended to divide their savings up evenly among them, the study found. This approach strikes University of Chicago behavioral science and economics professor Richard Thaler, PhD, as naive.

"Diversification is a good strategy; there is a reason why our mothers tell us not to put all our eggs in one basket," says Thaler. However, dividing your eggs evenly among available baskets is often not the best solution, in terms of investing.

"All heuristics get used because they make sense on average, and it's just when things get very complicated that they stop working," Thaler notes.

Failure to learn

Helping people discover what strategies do work may be an uphill battle, according to an additional finding in Rieskamp's study.

After students initially allocated their contributions, a computer simulation showed them how much interest they would earn that year. The students then had the opportunity to make adjustments as they contributed an additional $5,000 of their income into the 401(k). They did this 35 times, representing 35 years, but the participants did not improve as investors. Since stocks, in this simulation, tended to give participants a better return over time, many of them should have shifted their savings toward stocks. Instead they responded only to the most recent event-moving money out of a stock fund that did badly last year, without looking at how it performed over the course of the simulation.

"It is only the recent return that you use to decide whether the investment is one you should decrease or increase," Rieskamp says.

The finding fits with six decades of cognitive science research showing that people have trouble keeping track of trends when individual events have a lot of variability, notes Ido Erev, PhD, a behavioral science professor at the Israel Institute of Technology. When a stock fund shows a high return some quarters and finishes at the bottom of the pack in others, people seem to have a hard time estimating the fund's overall performance, he observes.

"Research shows that there are certain conditions when learning is very slow, and this is one of them," Erev says.

In fact, grocery store managers may make use of this phenomenon when they change prices on a daily basis. This makes it difficult for shoppers to recall which store tends to have, for instance, cheaper diapers on average, Erev notes.

So how can investors overcome this cognitive handicap? One solution increasingly favored by 401(k) providers is offering "lifestyle funds," which keep people's investments balanced among different kinds of securities, making recommendations based on the investor's age and risk tolerance, says Thaler. Such options are a good idea, he says, because they lead to smarter retirement investments, and ultimately, better retirements.

"This is probably the most important financial decision people make, other than who to marry," he says.