CHAPTER 1
THE PROBLEM: OUTLIVING YOUR RETIREMENT SAVINGS
Long-range planning for income from retirement savings is beset with many uncertainties. To begin with, we don't know how long we'll live in retirement. Will it be 20?years? Thirty years? Or will we survive to the ripe old age of 105, as did dear-departed Aunt Matilda?
Furthermore, financial markets are inherently volatile and unpredictable. We don't know what returns to expect from investments such as stocks and bonds in years to come. The future path of inflation is also indeterminate. Will the stock market crash, or will it soar? Will we return to the painful inflation of the 1970s, which mauled the portfolios of retirees, or will the Federal Reserve triumph in its battle to beat it back? As we'll see in later chapters, these are vital considerations for your withdrawal plan.
Long ago, I learned the futility of forecasting any of these factors. I prefer the attitude of one of my favorite authors, Mark Twain, who slyly observed, "Prediction is difficult - particularly when it involves the future." Thus, in the early 1990s, when I tackled the urgent need to guide my clients on retirement withdrawals, I decided to turn toward the past for answers.
1.1 PORTFOLIO LONGEVITY, BEAR MARKETS, AND INFLATION
My research relies heavily on charts to illustrate my findings. You'll find this book loaded to the brim with charts! An updated version of the first chart I ever produced, in 1993, appears in Figure 1.1.
This chart presents, for each of hundreds of historical retirees, the length of time their portfolio lasted (its "longevity"), assuming an initial withdrawal rate of 6%. The 6% rate was chosen arbitrarily. At the time, I had no idea what constituted a reasonable withdrawal rate.
Some clarifications are needed before we engage in observations about this chart.
First, this chart contains data for clients who retired on the first day of each quarter beginning 1 January 1926, and ending 1 January 2013, a total of 349 individuals.
Figure 1.1 Number of years portfolio lasted @ 6% initial withdrawal rate. Tax-advantaged account, cost-of-living adjustment (COLA) scheme, seven asset classes, capped at 50?years, fixed 55%/40%/5% stocks/bonds/cash.
The findings in this book are based on actual data through 31 December 2022. Analysis of retiree portfolios with less than 10?years of actual data can produce unreliable results, so for this chart I excluded individuals who retired after 1 January 2013.
Second, I capped the "longevity" of each portfolio at a maximum of 50?years. There are, of course, a few exceptionally long-lived individuals who will be retired for longer, and many of the "Financial Independence, Retire Early (FIRE)" folks (who aspire to retire at age 35 or younger) may exceed this limitation. However, 50?years is long enough for the great majority of retirees.
Third, those who retired on or after 1 January 1974 lack a full 50?years of actual data. For years with missing data (beginning 1 January 2023), I have inserted long-term averages of investment returns and inflation statistics. This "data gap" is not of great concern, as the longevity of most retirement portfolios is determined in their first decade. This doesn't imply that developments of the last 20?years of retirement are without impact; it's just that earlier events have their effects compounded for longer, and thus those effects are magnified.
Next, we need to specify some of the assumptions made in the preparation of the chart:
- A "tax-advantaged" account (such as an IRA or Roth IRA).
- The portfolio "rebalanced" (returned to its original allocation) once annually.
- Portfolio allocated as follows: 55% in stocks, 40% in bonds, and 5% in cash.
- Withdrawals made using the "COLA" scheme, which works as follows: The first-year withdrawal equals 6% of the portfolio's starting value. In subsequent years, withdrawals are increased by the prior year's inflation, much as Social Security functions.
Now that we have the preliminaries out of the way, let's take a closer look at Figure 1.1. One is immediately struck by how many portfolios, even at a seemingly high withdrawal rate of 6%, lasted for at least 50?years. More than a third of the retirees fall into this category. Many of these probably lasted considerably longer, but the chart is arbitrarily capped at 50?years. A "6% rule" would have worked for these retirees, and possibly even a higher withdrawal rate!
It's also evident that, in contrast, many retirees ran out of money well short of 50?years, including several who failed to make it to 20?years. These "failed portfolios" seemed to occur in clusters. What's the reason for that?
To answer that question, consider Table 1.1, which lists major bear markets in US Large-Company Stocks since 1926. A bear market is an extended period during which stock prices decline significantly. They are often associated with a recession in the economy. Furthermore, to be considered a bona fide major bear market, investors need to develop a negative psychology, often including a conviction that they "will never buy another stock as long as they live." By these standards, the 33% COVID-related decline in early 2020 does not qualify as a major bear market, as investors did not have enough time to develop FOLM (Fear of Losing Money). In fact, they barely had enough time to swallow their Paxlovid tablets before it was over.
Table 1.1 Characteristics of major bear markets in US large-company stocks (LCS) since 1926
Period Duration (months) Total return LCS Change in inflation Total return ITGB* September 1929-June 1932 34
-89.2% -27.8% +17.6% March 1937-March 1938 13
-50.0% + 0.0% +3.1% January 1973-December 1974 24
-48.2% +22.1% +10.6% September 2000-September 2002 25
-49.1% +4.8% +9.7% October 2007-March 2009 17
-56.8% +1.8% +15.0% * ?ITGB = Intermediate-Term US Government Bonds
Perhaps unsurprisingly, most of the dips in Figure 1.1 coincide with the major bear markets listed in Table 1.1. It's fair to conclude that a major bear market early in retirement tends to significantly reduce portfolio longevity. Concomitantly, a major bull market early in retirement (such as during the 1980s) can have the felicitous effect of extending a portfolio's life.
There are further lessons to be learned here. Note that the bar corresponding to the October 1947 retiree is notably short (22?years) but is not associated with any major bear market. However, from July 1946 through June 1948, a period of two years, inflation raged at high levels. This clearly indicates that inflation is a critical factor in making retirement portfolios last.
In Table 1.1, the deepest bear market occurred during the Depression, with a loss exceeding 89%. However, according to Figure 1.1, a more severe reduction in portfolio longevity happened a few years prior to the 1973-1974 bear market, which registered a much smaller loss in stock prices (-49.8%). Why didn't the Depression bear market have a more onerous effect on portfolio longevity?
The answer is that the Depression was a time of double-digit deflation (declining consumer prices), which reduced withdrawals for several consecutive years. The 1973-1974 bear market occurred during a period of high inflation (rising consumer prices), so that retirement portfolios were subjected to a double whammy of declining portfolio values and rapidly accelerating withdrawals (to better understand this dual effect, see the "balloon analogy" in Section 1.5). There have been other periods as well, when a moderate decline in stocks accompanied by a short burst of high inflation damaged retirement portfolios. In other words, inflation must be considered on a par with stock bear markets as a determinant of the success of a withdrawal plan.
Finally, the last column in Table 1.1 displays the returns earned by US Government Bonds during stock bear markets. The positive returns earned by bonds made them a useful "diversifier" to offset some of the losses in stocks. We shall see further evidence of this effect when we discuss optimum asset allocations for retirement portfolios.
1.2 MY RESEARCH METHODOLOGY
A few words about my research methods seem appropriate. From the beginning, I've used a...
