| Monte
Carlo Simulation: Smart Bet for Baby Boomers’ Retirement
Plans
By
Evan M. Levine
SEPTEMBER
2005 - The baby boom generation, those 77 million Americans
born during the postwar period of 1946 to 1964, has witnessed
the Dow Jones Industrial Average increase by a factor of
40 and has also enjoyed approximately 500% growth in home
equity during their adult lives. But somehow, 25 million
of the 77 million boomers currently have a net worth of
less than $1,000 (excluding the value of their home equity).
Boomers’ financial situation, combined with rising
life expectancies, means they will require high-quality
retirement planning advice. Many of them will want that
resource to be their tax advisor or a referral from their
tax advisor.
A CPA
firm that decides to provide this service can provide it
in-house by hiring a financial planner or a registered investment
advisor, or it can form a strategic alliance with another
firm of professionals. In either case, what is needed is
not just someone who executes transactions. The right person
is someone capable of providing accurate and forward-thinking
advice. Ideally, this will be an advisor who understands
the application of Monte Carlo simulations in retirement
planning.
The
Monte Carlo Method
Various
industries have used the Monte Carlo method for decades.
Its principles were first used in the 1940s by scientists
at Los Alamos, N.M., working on the atomic bomb. It has
more recently been applied to help urban planners predict
traffic patterns and to institutional investment portfolios
to forecast probable outcomes. Today, retirement planners
in retail financial services use it to adjust retirement
income planning for annual variations in projected “average”
returns. Consider the period 1968–1998, when the average
return for the S&P 500 was 11.7%. If a planner was advising
a client at the beginning of that period and actually had
the foresight to use 11.7% as an assumed average return,
it would appear that an individual could safely withdraw
8.5% of the portfolio’s initial value, then increase
withdrawals by 3% annually for inflation (see Exhibit
1). In reality, however, it wouldn’t have worked
out that way, because while the S&P 500 indeed averaged
11.7% per year, it did not deliver 11.7% each and every
year. Some years were much worse than the overall average
of 11.7% and other years were much better. Consequently,
had the individual followed the advice in this example,
she would have actually depleted all of her assets by 1981
because stocks performed very poorly in the first half of
this period (see Exhibit
2).
A Monte
Carlo simulation can help because, rather than rendering
advice using a flawed assumed average return for each and
every year of an analysis, an advisor can simulate the plan
using randomly ordered returns based on a set of reasonable
parameters. Computer software can simulate retirement cash
flows 500 or 1,000 times, thus reflecting a range of possible
outcomes (see Exhibit
3).
One
can then observe the results of a financial plan during
the peak of a bull market or during the trough of a bear
market. From this, a planner can arrive at a probability
of success: that is, how many times out of the 500 or 1,000
simulations the plan actually held up (i.e., the portfolio
assets were not depleted before the end of the time period).
Many planners look for somewhere between 75% to 90% success
to have sufficient confidence in a plan.
Because
this approach has its limitations, these models are only
as good as their assumptions, and using one certainly doesn’t
completely eliminate uncertainty. By recognizing uncertainty,
however, it is an improved, more sophisticated form of advice
as compared to traditional plans that use obviously erroneous
averages.
Useful
websites for learning more about Monte Carlo simulations
include www.financeware.com
and www.analycorp.com.
Evan
M. Levine, ChFC, is a financial advisor based in
Garden City, N.Y., specializing in retirement income planning
for baby boomers. He can be reached at 516-240-6161 or elevine@finsvcs.com.
|