There are many different strategies for withdrawing income from your investments.  If you own interest or dividend bearing securities and can live off that income alone, then odds are your financial security is secure (this article assumes you have made provisions for other possibilities such as rising health care costs, long-tern care and other hazards that can deplete your principal).   However, most retirees will not find this adequate, especially in years when the market may be down and as inflation takes it bite out of purchasing power.

The question then becomes how much of your principal you can afford to draw off and still provide for inflation, and not deplete your resources before you die.  Just a few short years ago, Peter Lynch, a well-respected mutual fund manager advised retirees that if they invested 100 percent of their money in common stocks, they could withdraw 7 percent a year (assuming a 10 percent average annual return) and still have lots of money.

The problem with that logic is that it just isn’t true.  Using an average rate of return (ROR) to project future income doesn’t work because your investments will not deliver that ROR each year.  If your early retirement years are marked by less than average, or worse yet, negative returns, you will reduce the principal that remains to grow.  This could result in you having to either reduce your income in latter years, which inflation will make difficult, or you could run out of money.  

Where in the time period you choose to retire, bad return years occur, impacts how long your money will last, but since that cannot be predicted in advance, it is important that you plan your withdrawal rate to allow for market fluctuations.

The Trinity Study [i]

Enter the Trinity Study.  This study was conducted by three professors, at Trinity University, a few years ago to study what withdrawal rates were least likely to deplete an investor’s funds, and how the composition of the portfolio, stocks versus bonds, impacted the withdrawal rate. The study looked at the impact of withdrawal rates that varied from 3 – 12 percent, on 5 different portfolios ranging from 100 percent stock to 100 percent bonds, over all rolling withdrawal periods of 15, 20, 25 and 30 years.  One of the important characteristics of this study is that it used real historical market data not average rates of return for those time periods.  It also took the effect of inflation into account and adjusted the withdrawal rates upward each year accordingly.

Revelations revealed

Contrary to the opinion espoused by Peter Lynch, this study found that over all 30 year time periods from 1946 through 1997, that a 100 percent stock portfolio would have been able to provide a 6 percent income, increasing each year with inflation, only 57 percent of the time.  This means that a senior who relied on a stock portfolio for this level of income ran out of money, before 30 years had passed, 43 percent of the time.  Given that life expectancies have risen, and many people retire earlier than age 65, this represents a sizeable risk.

The authors reached these five general conclusions:

1. Younger retirees who anticipate longer retirement payout periods should plan on lower withdrawal rates.
2. Bonds increase the success rate for lower to midlevel withdrawal rates, but most retirees would benefit from a stock allocation of at least 50 percent.
3. Retirees who desire inflation-adjusted withdraws must accept a substantially reduced withdrawal rate from the initial portfolio.
4. Stock-dominated portfolios using a 3 percent or 4 percent withdrawal rate may create rich heirs at the expense of the retiree’s current consumption.
5. For 15-year or less payout periods, a withdrawal rate of 8 to 9 percent from a stock-dominated portfolio appears to be sustainable.

See for yourself

Here are the links to the 4 tables of data created by the study.

• Table 1 illustrates the success rate of various portfolios for different time periods measured against the full time span of the Ibbotson data used, 1926 – 1995.
• Table 2 illustrates the success rate of various portfolios for the time period after WW II, from 1946 – 1995.  As expected, market returns were better during that period and thus success rates improved significantly.
• Table 3 illustrates the success rate of various portfolios for different time periods, adjusting for inflation/deflation during the period.  As you might expect, this one provides the gloomiest results.
• Table 4 illustrates the variations in the amount of money you might have left at the end of each time period.  Remember, that the success rates of the above tables only show what portion of time a given withdrawal rate avoided depleting the portfolio.  In fact, each different time period produced a different result, and the range is enormous.  This table best illustrates the risks of using a static assumption for ROR.