Asset Allocation for the Ages.

This article was posted on September 1, 1998.

The financial turmoil in Asia and Russia caused a "flight to safety" in August, pushing the yield on the 30-year US Treasury bond to an all time low of 5.38%. Should investors with a long term outlook jump on this bandwagon? Are bonds really safer?

Many academic studies have revealed that asset allocation is the most important factor in predicting a portfolio's investment return. Asset allocation also determines the overall volatility or "risk" of a portfolio. Most investors, and especially retirees, seek to maximize return and minimize risk.

Asset Allocation for the Ages or "The Trinity Study" meets "Stocks for the Long Run."

The two most influential retirement planning works I've read and used are Stocks for the Long Run, by Jeremy J. Siegel, Professor of Finance, the Wharton School of the University of Pennsylvania and Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable, by Philip L. Cooley, Carl M. Hubbard and Daniel T. Walz. (AAII Journal, February 1998, Page 16) The latter work is popularly know as the "Trinity study" since the three authors are professors of finance at Trinity University in San Antonio, TX. The "Trinity study" has also been promenently mentioned in Scott Burns' columns in the Dallas Morning News and Worth magazine.

The surprising conclusion of both texts runs contrary to the "conventional wisdom" that investors should switch to bonds once they reach retirement age. For investors with 25 or 30 year payout periods, portfolios with more than 30% bonds have more volatility and lower total returns than more stock heavy asset allocations. That's right. These studies reveal you can actually increase your chances of going broke by selling stocks and buying bonds.

The table below from Stocks for the Long Run used the concept of efficient frontiers to determine the asset allocation with the least risk for holding periods of varying lengths. Bonds and cash are appropriate for 1 to 5 year periods. A compelling case can be found for stocks when the holding period is 10 years or more.

The most interesting insight from the "Trinity study" is that portfolios with a substantial allocation of stocks are more likely to "survive." The plot below shows various allocations of stocks and bonds for a portfolio with a 30 year payout period. The portfolio success (i.e., the probability that the portfolio asset value will not drop to zero during the payout period) is higher for stock dominated portfolios at almost every "inflation adjusted" withdrawal rate. The fact that the withdrawals are "inflation adjusted" illustrates the most conservative presentation of the data. The decrease in the "buying power" of a retiree's income with inflation and the passage of time is often the largest threat to a retiree's standard of living. Retirees seeking a 100% probability of success must keep their initial withdrawal slightly below 4% of assets, then increasing the annual withdrawals for inflation in subsequent years. See, "What's the maximum safe withdrawal rate in retirement?"

What about inflation adjusted bonds ?

The US Treasury now offers Treasury Inflation Protected Securities (TIPS) in 5, 10, and 30 year maturities. Wouldn't these securities offer the best insurance against inflation and the risks of the stock market?

TIPS certainly offer insurance, but at a very steep premium. On August 21, 1998 the yield on the 30 year TIPS was 3.74%. The August 10, 1998 issue of Barron's reported that worst 30 year period for the S&P 500 since the end of World War II (1955-1984) had an annual return of 9.4%, the best 30 year period (1968-1997) had an annual return of 12.1%. That's a pretty narrow range for the best and worse case. Adjusting the worst case return of 9.4% for the 4.7% average inflation (measured by the CPI) during that period would give you a 4.7% inflation adjusted return. Forgoing 20% of your projected return (4.7%-3.74%=0.96%, 0.96%/4.7% = 0.20 = 20%) to protect yourself against an event that hasn't occurred in recent history is very expensive insurance. You would likely have a greater probablility of collecting on an insurance policy protecting you against being killed by an asteroid or being abducted by aliens.

What's the most reasonable interpretation of these results ?

These studies offer several points that should be useful to retirees in their investment planning:

• Asset Allocation depends on your "payout period", not your age. The average life expectency of a person age 60 is 25 years. A retiree in his 80's planning on leaving the bulk of his investment portfolio to his grandchildren might have a 50 year payout period.
  
  On the other hand, an 80 year old planning to "spend it while I'm alive", might very well have the less than 10 year payout period that would favor a portfolio of bonds and cash. The lesson here is to look at "payout period" rather than age.
  
  Applying the concept of payout period to my own case, I've put 1 to 2 years worth of living expenses in a money market fund, 5 years worth in a laddered portfolio of 1 to 5 year Treasury securities, and the balance in a diversified portfolio of domestic and foreign stocks. In effect, I've decided to limit my equity exposure to money I won't need for at least 6 or 7 years.

• The concept of "payout period" has implications for other investment needs. You can also apply this concept to saving for college or a down payment on a home. For example, a family saving for a child's college education would want the money to be primarily invested in stocks until the child reached age 13. Anticipating tuition bills at age 18 to 22, the family would want to start moving these funds to bonds and cash about 5 years in advance of the first tuition payment.

• Eliminating the last sliver of risk from your financial life has a high cost. Inflation insurance in the form of inflation-indexed bonds is expensive. There is a very high probability that stocks will beat inflation indexed bonds for investors with 20 to 30 year payout periods. As Professor Siegel writes, "the last 30-year period when bonds beat stocks ended in 1861, at the onset of the U.S. Civil War." Only the very wealthy, or the most risk adverse, should consider giving up a large portion of their projected investment return on a bet that this unlikely event will reoccur.