Stanford University Home

Stanford News Archive

Stanford Report, April 15, 1998

Economists seek to improve financial advice: 4/15/98

Economists seek to improve advice on investing pension money


"We'd all like to see a higher rate of return on . . . investments. The question is how do you get that and still keep the [Social Security] system that has lifted so many seniors out of poverty and dealt with disability and premature death."

­ President Clinton, April 7,
Kansas City, Mo.

As a debate begins about whether Americans should invest their own Social Security dollars in private markets, evidence mounts that workers already face hardships getting sound advice about investing their private pension dollars.

Two Stanford researchers who are trying to improve that financial advice are economists John Shoven of the Economics Department and William Sharpe of the Graduate School of Business.

Related Information:

Sharpe, the STANCO 25 Professor of Finance who won a Nobel Prize in 1990 for his work on reducing the risk of investments, heads a new company that hopes to provide individuals with the same quality of personalized investment advice that he and others already provide to billion-dollar pension fund managers. Joseph Grundfest, the W.A. Franke Professor of Law and Business, and a former commissioner of the Securities and Exchange Commission, also is a principal in the company.

Shoven, an expert on national savings, Social Security and retirement planning who is the Charles Schwab Professor of Economics and dean of the School of Humanities and Sciences, also analyzes the risk-return tradeoffs involved in individual retirement planning. By carefully picking scenarios that broadly apply to high- and middle-income households, he tries to provide general insights into the changing retirement planning picture. His considerations include changing tax laws as well as market conditions.

Shoven's most recent study, "The Location and Allocation of Assets in Pension and Conventional Savings Accounts," was published by the Stanford Center for Economic Policy Research and presented April 3 at a Wharton School of Business conference. It finds that some of the conventional wisdom on asset allocation may be too conservative and, in some cases, even the opposite of what it should be.

Getting sound advice out to millions of investors has become vastly more important since 1978, when Congress broadly authorized 401(k) plans. In such plans employers and their employees usually contribute to an individual retirement portfolio chosen by the employee rather than by a professional fund manager. About 80 percent of American workers now invest a total of $1.5 trillion in such plans.

Shoven and Sharpe both use computer technology and advances in theory to provide information beyond the reach of most personal financial planners, let alone individual investors. A legitimate complaint about the analysis many advisers and computer programs provide, they say, is that they assume stocks will earn annual returns of a particular percentage and bonds will earn at another slightly lower rate. In the real world, returns fluctuate, creating uncertainty.

Independent companies rate the past performance of mutual funds and newspapers regularly publish their current performance, but the performance of a particular fund is not what is important to an investor who has more than one fund. What counts is the overall performance and combined risk of the investments. Two investments that tend to go up or down in value at the same time, such as Ford and Chrysler stock, or German and Dutch currency, are more risky to hold than two whose values tend to change independently, but such correlations are not always this obvious to an investor.

Modern portfolio theory, for which Sharpe and Harry Markowitz shared a Nobel Prize, both recognizes the uncertainty of returns and computes the historical covariances between holdings. It uses rigorous mathematical analysis to maximize the chances of achieving an optimal return on investments, given an individual's' retirement income goal and his or her willingness to accept the risk of undershooting the goal.

Shoven uses a computer to generate 10,000 random 30-year experiences with a particular pension saving strategy. He compares the strategy's mean performance to the results of similar simulations for another savings strategy. But since not everyone will have the average outcome of 10,000 outcomes, he also looks at the outcomes at other percentiles to answer the question, what happens if you are unlucky a little bit or a lot?

A risk-averse person, he told a recent campus gathering, might care more about the outcome at the 5th percentile ­ what would happen if he or she were unlucky enough to do worse than 95 percent of the outcomes expected from a particular investment strategy. A less risk-averse person might focus on the 25th percentile. Accepting a smaller likely return in exchange for less risk, he explained, can be thought of as "buying insurance," which in some cases can be quite expensive.

Shoven's most recent simulations reached these conclusions: Even a highly risk-averse person would be prudent to put 60 percent of his or her pension savings in stock mutual funds, rather than in bonds, because he or she would earn at least as much as with more bonds even at the worst outcomes. Someone willing to tolerate somewhat greater odds of a downside should consider 70 to 90 percent stocks.

Shoven's simulations are generalized to a worker saving the same amount for 30 years for his or her retirement. But even someone who is only 15 years away from drawing out his or her retirement income would likely do slightly better to invest in more stocks than bonds, he found in further simulations.

Higher income households who invest outside of tax-deferred pension plans as well as within them can make substantially more, he concluded, by putting their stock investments inside the pension and municipal bonds outside it, rather than following the more conventional advice, which holds that the asset facing the highest tax rates should be inside a tax-deferred pension plan. The reasons, he said, relate to the different tax treatments of investment categories held outside a pension, the low implicit tax rate on municipal bonds compared to corporate bonds, and the fact that stocks, especially most stock mutual funds, produce a steady flow of capital gains. Capital gains earnings are taxed annually outside the pension, but taxes on them are deferred when held within, he said.

Sharpe views Shoven's analyses as "a step in the right direction. " It provides investors with more meaningful information about the importance of diversification and the tradeoff between risk and return. But, he says, the falling cost of computer power now makes more customized advice possible. His company, Financial Engines Advisers of Palo Alto, is "pushing the envelope of both financial theory and programming," he said last week, in hopes of delivering customized analysis to individuals over the Internet later this year. The service would be available at first only to employees of employers who become clients. (Stanford has not signed on as a client, but there have been some discussions between the university's Benefits Office and his company about that possibility, Sharpe said.) A pilot service is currently running in four companies.

"You couldn't do this on the scale we hope to do it and with the kind of costs we hope to reach before Java [programming language] and the Internet," he said. "If we were to do this much work for each person on a server, it would either cost too much or when a peak load hit, the system would die."

Once such data are obtained, however, there is still an art involved in presenting it to individuals, Sharpe said. "Ordinary human beings get overwhelmed when they look at a whole distribution curve" of possible outcomes. The shape of the curve partly determines how many numbers people need to understand in order to evaluate the risk-value tradeoff of a given strategy, he said. "Generally they need to focus on two numbers in order to take into account both return (or opportunity) and risk (or the results associated with bad luck)." Depending upon what the person thinks of these results, a new strategy can be developed and analyzed that will likely push one number up and the other down, but occasionally both numbers get better or worse together.

Such "rocket science money management," as the Wall Street Journal called Sharpe's model, could substantially change financial advising services and may influence the decisions the nation makes on changing its Social Security system.

President Clinton, who spoke at the first of four regional forums on Social Security reform last week, made it clear he is open to the idea of allowing workers to invest all or part of their Social Security taxes in private markets. Two members of Congress who were with him backed the idea of individual private accounts, a third said it needed more study, and a fourth favored pooling the funds for private investment. Clinton said he will make a formal proposal on Social Security reforms at a White House conference in December. SR