These are unsettling times for retirees who depend on their investments for income, but the right strategy can help tremendously with the preservation of assets.

With markets erratic, many older people may be tempted to dump stocks. But that would be a mistake, according to a study in the

Journal of Financial Planning

. The study says that older people should hold onto their stocks and sell bonds first to cover living expenses.

"The higher your stock allocation, the less likely you are to run out of money," says John Spitzer, a professor at the

State University of New York -- Brockport

, who conducted the research along with a colleague, Sandeep Singh.

The study flies in the face of conventional wisdom, which holds that older people should increase their stakes in bonds. Still, the findings are worth considering, at a time when managers of popular target-date portfolios and other funds are working hard to find the best mixes of stocks and bonds.

In recent years, some of the funds have altered their thinking, changing to bigger stock allocations. If that shift continues, the professors -- who now are on the fringe -- could become close to the mainstream of the investment world.

To appreciate the recent study, consider this: For decades, financial advisers have held a central belief about portfolio management. As investors age, the thinking goes, they should reduce stocks and increase the percentage of holdings in bonds.

The target-date funds

follow this line of thought

. Holding broad mixes of stocks and bonds, the funds are designed for savers who plan to retire in particular years, such as 2020 or 2030. As the retirement date approaches, the funds lower their equity allocations and increase bond holdings.

An investor in

Fidelity Freedom funds

would have about 81% of assets in equities at a point that is 23 years before the retirement date. The figure drops to 43% two years after retirement, and 20% 15 years after retirement.

The target funds offer convenient diversification for millions of investors who never had it before, and shareholders seem to be embracing this approach

even if the funds may not be perfect for everyone

. Assets in the target funds climbed from $114 billion in 2006 to $182 billion in 2007, according to the Investment Company Institute, the mutual fund trade group.

Examine the Test

To test whether target-date and other conventional portfolios actually provide protection, the two professors looked at thousands of scenarios, examining how different strategies would have performed under historical market conditions.

In one scenario, a retiree has 50% of assets in stocks and 50% in bonds. The investor withdraws 4% of assets annually. To raise cash, the investor begins by selling bonds. Once the bonds are exhausted, the saver sells stocks to take withdrawals. In other scenarios, the investors sell stocks first or a mix of stocks and bonds.

For each scenario, the professors calculated whether there would be a shortfall -- an instance when the portfolio would have been exhausted.

In the study, the strategy of withdrawing bonds first was the clear winner, producing the fewest number of shortfalls. Professor Spitzer explains why: In the portfolio with 50% in bonds, the saver would spend his fixed-income holdings in about 12 years. During that time, the stocks would produce returns and gradually account for 100% of the portfolio. Since stocks generally outdo bonds, the investor would likely have more money in the stock-heavy portfolio than in one that has a mix of stocks and bonds.

Fund managers don't dispute Spitzer's math, but they say that the approach is too risky.

Spitzer concedes that in the real world, an all-stock portfolio could present problems for many retirees. As the markets have demonstrated recently, stocks can be volatile, bouncing sharply up and down. Portfolios that include stocks and bonds would deliver a smoother ride.

"Many investors may feel more comfortable with a portfolio that includes at least 20% in bonds," Spitzer says.

Fully aware of the risks posed by stocks, managers of target-date funds have taken a variety of approaches. Funds run by

Oppenheimer

and

T. Rowe Price

have relatively heavy allocations to stocks, while Vanguard has more weighting in bonds.

"There is no easy solution to the problem of asset allocation," says T. Rowe Price portfolio manager Jerome Clark. "We have focused on trying to avoid shortfalls, while other companies emphasize reducing volatility so that retirees don't have a sudden loss in a bear market."

In recent years, the stock-heavy approach has gained more support as Fidelity and Vanguard have increased their stock weightings.

In 2004, Fidelity's fund for savers with 10 years to go before retirement had 48.6% in equities. Today the comparable fund has 67.4% in stocks. Fidelity said that it made the change because shareholders were living longer and needed more equities to cover the expenses of long retirements.

"The funds evolve as demographics change and we learn more about different investment strategies," says Fidelity portfolio manager Jonathan Shelon.

Fidelity and the other managers of target-date funds are engaged in ongoing research, seeking to fine-tune their asset allocations. Shareholders who buy target funds now should recognize that the asset allocation strategies may not be carved in stone. Over the years, the stock weighting could increase -- and that isn't necessarily a bad thing.

Stan Luxenberg is a freelance writer who specializes in mutual funds and investing. He was formerly executive editor of Individual Investor magazine.