After twenty-five years of distribution, the impact of the first-year
allocation has increased the ending value from $94,532 to $264,191
(a 179-percent increase). Seemingly insignificant tactical adjustments can have a big impact on distribution portfolios. This example shows how such an adjustment can have a positive impact. But
without the safety net of time, the wrong adjustment can have a
significant negative impact.
The butterfly effect also applies to fees. Table 6 shows ending values
after twenty-five years of accumulation and twenty-five years of distribution for two investments that are identical except that the first
has annual fees of 1. 5 percent and the second has annual fees of
1 percent. The difference in fees has a significant impact on the ending accumulation values ($3,322,256 versus $3,743,020) but a critical
impact on the ending distribution values ($95,865 versus $312,650).
Although fees have an impact, it is also important to remember
that price is what you pay but value is what you get. When working
with a smaller margin for error, effective risk management is more
important than fees—and the butterfly effect applies to risk, cash
flow, and fees. Retirees who choose an advisor to manage their life
savings based strictly on fees could find themselves going broke
cheaply, as shown in figure 4.
Figure 4 shows year-by-year account values for two portfolios, both
with 5-percent initial withdrawals that increase 3 percent annually.
The black line shows the results of using a less-volatile risk alloca-
tion (20/80) with a higher annual fee ( 1. 5 percent), and the gold
bars represent a more volatile allocation (80/20) with a lower fee
(0.5 percent). By losing less early, the 20/80 allocation finished with
greater wealth ($565,074 versus $315,306) despite earning a lower
return ( 5. 1 percent versus 5. 4 percent) and paying a higher fee
( 1. 5 percent versus 0.5 percent). Taking this to the extreme, some-
one with an even more-volatile allocation (i.e., 100-percent
S&P 500 Total Return Index) but even lower fees (i.e., no fees)
would have had the highest average annual return ( 5. 6 percent)
but finished with the least wealth ($167,155). Thus, someone could
have paid 1. 5 percent per year for a portfolio (20/80) that did not
beat the market (the S&P Index) yet was considerably more success-
ful (finishing with 56 percent of investment versus 16 percent). The
bottom line is that retirees who focus on just fees or just beating an
index risk winning a battle but losing the war.
Systematic Approaches Are Inefficient and Dangerous
Systematic investment is the type of commitment device that has
helped many save for retirement. But a systematic approach to
income post-retirement is inefficient and may be dangerous.
Consider the example in figure 5, which shows results for an investor who retired in 1965 with a 50/50 allocation and withdrew
5 percent the first year. After twenty-five years, the systematic
approach of increasing withdrawals by 3 percent annually (gold)
left the investor with $79,192 ( 7. 9 percent of the original investment). If the investor had kept the year-one distribution level for
year two, then increased the cash flow by 3 percent every year
(blue), the ending balance is $242,655. If the investor doesn’t
increase income in any of the five years during which the portfolio
had a negative return (green), the ending balance is $626,399.
Industry research focuses on the impact of the initial withdrawal,
but the frequency and amount of subsequent increases are as
important if not more so. Staggering the withdrawal increases
improves portfolio longevity and systematic models significantly
overstate longevity risk.
Psychology: What’s the Worst That Can Happen?
Decision-makers intuitively rely on rules of thumb known as
“heuristics” that simplify complex mental tasks but may lead to
suboptimal decisions (Tversky and Kahneman 1974). By assuming
that investors always act rationally, traditional portfolio management fails to account for these unseen drivers of decision-making.
Loss aversion is one bias that changes significantly post-retirement
and may make retirees more prone to panic. For the average person,
small to moderate losses have twice the emotional impact as equal
Figure 5: Systematic versus Managed Withdrawals Figure 4: Risk and Fees
Year-by-Year Account Values 2000–2013
$1m Investments, 5% Withdrawals with 3% Annual Increases
Ending Values 1965–1989
$1m Investments (50/50) with 5% Withdrawals
2000 2001 2002 003 2004 2005 2006 2007 2008 2009 2010 2011 2012 013
80/20 (0.5% fee) Averaged 5.4%
20/80 ( 1.5% fee) Averaged 5.1%
Systematic Year 2 5 Negative Years