This illustrates why much of the industry promotes 4 percent as a
safe initial withdrawal, leading to “the 4-percent rule.”
These outcomes, however, are based on accumulation-style thinking regarding volatile static portfolios, systematic approaches, and
the flawed assumption that retirees must increase cash flow 3 percent annually to maintain standard of living. These assumptions
overstate inflation and longevity risk and understate the value of
active management. As Scott et al. (2008) noted, trying to finance a
constant spending plan with a volatile investment is fundamentally
flawed because it will result in shortages in some environments and
surpluses in others.
The same flawed assumptions around static growth portfolios,
systematic withdrawals, and inflation, coupled with misalignment
with retiree psychology, limit the practical application of probability forecasting as a retirement-income tool. Such forecasting tends
to focus on the probability of an event, but an availability heuristic
causes retirees to focus more on the magnitude of a high-impact
low-probability event, specifically the recent historical worst-case
scenario. In other words, retirees do not care if solution providers
run a thousand scenarios, they will live through only one, and they
want to have a plan should it be a worst-case scenario. Hence the
terms “probability” and “Monte Carlo analysis” generate a preponderance of negative responses from investors. Probability forecasting significantly overstates longevity risk and fails to account for
the positive impact of active management of risk and cash flow
(see table 5 and figure 5).
Adaptive Distribution Theory
Innovation and adaptability are the primary drivers of survival in
financial markets (Lo 2005). This is especially true in the case of
retirement income, where you must account for multiple goals, risks,
and behavioral biases that change across constantly shifting markets
—and there are no second chances. Accounting for so many moving
parts requires comprehensive vision and adaptability, yet most
research is based on flawed assumptions, old strategies, narrow
framing, and systematic or static rather than adaptable approaches.
An alternative approach I call “adaptive distribution theory”
creates a personalized cash-flow plan by balancing multiple goals,
accounting for behavioral biases, redefining risk, emphasizing
short-term strategy, and assuming the worst then adapting.
Adaptive distribution theory addresses the cornerstones of risk and
cash-flow allocation by managing to acceptable annualized erosion
rates and building a buffer to absorb risk.
What Retirees Want: The 4M’s
A better understanding of what retirees want can lead to a more-
personalized solution. Retirees want to do the following:
• Maximize current cash flow, especially in the early years of
• Maintain their standard of living (and they can do so with much
less than 3-percent annual increases).
• Minimize cash-flow risk.
• Minimize principal erosion, short- and long-term.
The key to comprehensive retirement income is seeing how
these goals are connected and balancing them to create personalized solutions.
For example, figure 9 shows the ending account values when you
apply a 5-percent initial withdrawal to the same time period as
figure 8 but relax the assumption of an annual increase and instead
use variable annual rates of increase. The first bar corresponds to
the line in figure 8, which left the investor with approximately
9. 5 percent of principal, too close to depleting savings. Successive
bars in figure 9 show the ending values with variable increases.
Figure 9 shows nine approaches to personalized solutions for
retirement distributions, each reflecting a different combination
of the 4M’s. It shows how someone who wants to maximize initial
cash flow could have exceeded the industry-standard 4-percent
initial withdrawal, increase the annual withdrawal less often, and
still minimize principal erosion without any unexpected reductions
in cash flow.
For instance, by increasing cash flow every other year (“ 1 of 2”) the
investor could have withdrawn 5 percent initially and finished the
twenty-fifth year with 56 percent of the original investment, likely
an acceptable level of long-term principal erosion for some retirees.
Figure 9: Staggered Increases in Retirement Withdrawal
“In other words, retirees do not care if solution providers run a thousand
scenarios, they will live through only
one, and they want to have a plan
should it be a worst-case scenario. ”
Ending Account Value 1957–1981
50/50 Allocation, 5% Initial Withdrawals with Different Frequencies
Years Withdrawals Increased
EveryYear 4of5 3of4 2of3 1of2 1of3 1of4 1of5 NoIncrease