Authors: Douglas
B. Bernheim,Jagadeesh
Gokhale
and Laurence
J. Kotlikoff
Publisher: MIT
Press
(1)
Reviewers:
Sherman
D. Hanna and Kathy Zink
Consumer
and Textile Sciences Department, The Ohio State University
Traditional
financial planning software asks what you would like to have in retirement.
ESPlanner claims to be the first software to help a household decide how
much is feasible in retirement. Actually, the Life Cycle Savings program (Hanna,
1989; Hanna, 2000; Hanna, Fan & Chang, 1995) has been used for 10 years
at Ohio State University and some other universities, but the authors’s
claim is reasonable in terms of commercially marketed software. The program
ESPlanner was developed by three
economists who have been active in research related to household savings
and retirement adequacy. The ESPlanner web page offers valuable insights,
and is worthwhile examining even by those who have no plans for using the
program. The following description is on the web page:
ESPlanner
is based on the fundamental goal of saving and insurance – the desire to
avoid major disruptions in your household’s living standard through time.
This “smoothing” of your living standard isn’t easy. ESPlanner uses advanced
mathematical techniques to calculate the saving and life insurance you
need to balance consuming in the present with consuming in the future and
to preserve your household’s living standard for survivors.
By
following ESPlanner’s annual recommendations, you save more when you can
and less when you can’t. Your saving, not your life style, adjusts to your
economic circumstances. Your life insurance holdings also change with changes
in your insurance needs.
ESPlanner
takes into account all of your economic resources, tax liabilities, and
Social Security retirement, survivor, and other benefits. It considers
your family composition, tax-deferred saving, housing plans, special expenditures,
estate plans, capacity to borrow, life-style preferences, and other key
factors.
ESPlanner’s
contingent planning recognizes that survivors may have special needs and
different incomes. ESPlanner also lets you quickly vary key inputs, like
your ages of retiring, collecting Social Security benefits, and withdrawing
tax-deferred assets, to determine how these choices alter your maximum
sustainable living standard.
Readers
who have used the Life Cycle Savings program will recognize the consumption
smoothing idea. It is in fact a fundamental starting point for rational
planning for savings. For instance, rather than have some arbitrary goal
for how much to accumulate for retirement, or even for how much one wants
to spend in retirement, both ESPlanner and the Life Cycle Savings program
start with initial net worth and projected income from jobs and defined
benefit pensions, and calculate feasible levels of spending over the remaining
lifetime. The difference is that ESPlanner is much more complex because
it tries to accomplish much more. In classes at Ohio State, we get most
students ready to run the Life Cycle Savings program after a one hour lecture.
It is difficult to imagine that being possible with ESPlanner.
Consider
the information required from the household for ESPlanner, as listed on
the web page:
a.
Pay stub for you (and your spouse.)
b.
Ages you (and your spouse) plan to retire.
c.
Projected employee and self-employment labor earnings through retirement.
d.
Projected special expenses on college, weddings, emergency funds, airplanes,
nursing home care, and loan repayments.e. Projected special receipts
from bonuses, inheritances, sale of business, etc.
f.
If you rent your principal or vacation homes, monthly rent and other monthly
rental expenses.
g.
If you own your principal or vacation homes, for each home, its estimated
market value, annual property taxes, annual homeowners insurance, annual
condo fees and maintenance expenses.
h.
For each mortgage/loan for each home the outstanding mortgage balance,
the monthly mortgage payment (excluding escrow payments for property taxes
and homeowners insurance), and years left to pay.
i.
Projected funeral expenses and desired special bequests.
j.
Face and cash values of all individual, employer-provided, and group life
insurance policies on you (and your spouse)
k.
Balances in non tax-favored (non-retirement) accounts.
l.
Balances in IRA, Roth IRA, Keogh, SRA, and other tax-favored (retirement)
accounts. If married, separate balances are needed for each spouse.
m.
The amount you expect to add or subtract this year from non tax-favored
accounts. (ESPlanner uses this information in helping determine how much
you are actually saving.)
n.
Projected annual contributions to tax-favored accounts (retirement accounts)
by you (and your spouse.)
o.
For each defined benefit pension you (your spouse) will receive the year
it will start, the amounts of lump-sum and annual benefits, the degree
of inflation indexation and survivor benefit protection.
p.
Social Security earnings records for you (and your spouse).
Many
of these items are also necessary for running the Life Cycle Savings program,
although it is not clear why a program that does consumption smoothing
needs to ask for how much one is already saving – the point is that the
program will inform the user how much should be saved to reach goals.
ESPlanner
also asks for inputs on a number of obscure items, such as how much one
wants the household standard of living to increase or decrease over time.
The Life Cycle Savings program does not give a direct choice on that, because
it relies on assumptions about the risk of death, planned changes in household
size, the nature of the intertemporal utility function, and the real rate
of return faced by the household – all in a black box format.
ESPlanner
also asks for a parameter of economies of scale in shared living. All this
is very nice, except that it further complicates a very complex program.
ESPlanner
uses some unconventional terminology – for instance, one of the 28 different
reports the program can generate is called the “Household Non Tax-Favored
Balance Sheet.” While this report does include a column for non tax-favored
net worth, five of the six columns with dollar amounts are for annual amounts,
including income, taxes, spending, and saving. In another report, “household
consumption” does not include expenditures for housing or life insurance.
Even though these distinctions have some basis in economic theory, they
are not conducive to ease of use.
We
attempted to compare ESPlanner to the Life Cycle Savings program with one
scenario. It impossible to have a complete comparison, because ESPlanner
provides many more analyses than the Life Cycle Savings program. For instance,
ESPlanner suggests amounts of life insurance to purchase on each household
head.
It
is also difficult to ascertain whether we ran ESPlanner correctly. We spent
several hours on the project, but perhaps not enough. The scenario we ran
had a couple, each age 28, who planned to retire when they turned 67. They
planned to have two children and provide part of their college costs. Initially
the husband’s gross salary was $40,000 per year, increasing in real terms
to $116, 626 just before retirement. Initially the wife’s gross salary
was $30,000 per year, increasing to $59,492 just before retirement. They
planned to buy a $100,000 home at age 30, making a $20,000 down payment.
We
were comforted by the fact that both programs had similar projections of
Social Security pensions. However, the Life Cycle Savings program suggested
saving less than $2,500 the first year, compared to $26,247 recommended
by ESPlanner. ESPlanner suggested buying $1,174,333 of life insurance for
the husband the first year, and $1,290,597 for the wife.
There
is a newer version of ESPlanner now, so it is possible that these unrealistic
life insurance results were due to a bug in the 1999 version of the program,
or to a mistake by us. Despite the higher amount recommended for savings
each year, ESPlanner’s projected accumulation of financial assets was not
much higher than the amount projected by the Life Cycle Savings program.
We know the assumptions of the Life Cycle Savings program, and the real
rate of return assumed on investments for this comparison was 4% per year.
We are not sure about the assumptions used for ESPlanner. Both programs
carry projections out to age 100. Unless otherwise specified, ESPlanner
will try to use up all funds by retirement, leaving only enough for a funeral.
For this scenario, the Life Cycle Savings program projected that $118,696
(in constant dollars) would be left if both the husband and wife lived
to be 100 and did not have any extraordinary expenses such as long term
care. (It is possible to specify a larger amount.)
I
do not think ESPlanner would be good in most financial planning classes,
because the consumption smoothing aspect can be done much more easily with
the Life Cycle Savings program, and the extra features of ESPlanner, such
as estimating life insurance needs, can be done by students with more learning
benefits (e.g., Hanna, Gutter & Gibbs, 2000).
The
ESPlanner web site lists some financial planners who use the program. If
any of those planners are using it on a consistent basis, that would be
a great endorsement.
References
Hanna,
S. (1989). Optimal life cycle savings. Proceedings of the Association
for Financial Counseling and Planning Education, 4-12.
Hanna,
S. (2000). Life cycle savings program.. [WWW] www.hec.ohio-state.edu/people/shanna/lcs/overview.htm
Hanna,
S., Fan, J. & Chang, R. (1995). Optimal life cycle savings. Financial
Counseling and Planning, 6, 1-15.
Hanna, S., Gutter, M. & Gibbs, R. W. (2000). Life insurance math.
in Garman, E. T. & Xiao, J. J. (eds.)The mathematics of personal
financial planning, Second edition, 256-280.
How
to Avoid Outliving Your Money
Author: Moshe
Arye Milevsky, Chris Robinson and Kwok Ho
Publisher: Captus
Press
(2)
D. Hanna
Consumer
and Textile Sciences Department, The Ohio State University
The
web site for this software, which is an Excel template with many macros
for complex calculations behind the scenes, gives a detailed description
of the software.
Financial
Planners, take advantage of this revolutionary software that helps you
understand and explain:
The
effect of different asset allocation strategies on how much money one needs
to retire;
The
effect of risky rates of return and uncertain time of death on the probability
of experiencing a shortfall in retirement income;
The
effect of financial asset allocation strategies and time-to-retirement
on the probability of reaching a desired retirement goal.
You
can use How
to Avoid Outliving Your Money to better inform your client’s or your own
financial planning. It can help you develop a financial plan leading to
and during retirement.
This
software uses an advanced stochastic model and displays the effects of
uncertain life expectancy, given a person’s current age and gender (using
mortality tables) and risky rates of return. You can use historical averages
of long run returns from different asset classes that the software provides,
or specify your own expectations. No other personal finance software allows
for these uncertainties in such a rigorous fashion. Prof. Milevsky bases
the software on a series of award-winning academic papers. Now you can
access this expertise in a user-friendly format!
I
discussed this approach to retirement planning in my review of a personal
finance textbook written by some of the same people who developed this
software (Hanna, 1998). The basic issue is that financial planners should
take uncertainty into account in investing to reach goals, whether saving
for college (Hanna & Chen, 1996) or deciding on portfolio allocation
and prudent withdrawal rates after retirement (Cooley, Hubbard & Walz,
1999). Some analyses, such as those just listed, use a non-parametric analysis
based on past time periods, e.g., what would have happened in the worst
20 year period since 1926. Other analyses, such as this software, assume
some mathematical distribution of returns, and project the future. The
main difference is that theoretically worst long term period could be far
worse than we have observed. For example, we have not observed a 20 year
period since 1926 where an investment in an S&P 500 index fund would
have lost money, even in real terms, but a mathematical projection would
allow for that possibility.
Milevsky,
Robinson and Ho
have developed a slick spreadsheet that is relatively easy to use, and
will probably provide a reasonable answer to the question of how one can
set one’s portfolio allocation and withdrawal rate to obtain as much as
possible without having too high a risk of outliving one’s money. Do-it-yourself
investors might be satisfied to peruse the retirement section of the Motley
Fool web site (www.fool.com). Personally, I expect to keep most of my portfolio
in stocks, and with a defined benefit pension providing for fixed expenses,
I am willing to ride up and down with stock market returns, even if they
do not match the 1982-1999 results. However, a financial planner needs
to be more cautious. A client might expect to live only 15 years past retirement,
but it is possible that the client might live 35 years in retirement. Milevsky,
Robinson and Ho‘s
program allows for that possibility.
Even
though the spreadsheet is easy to use, I have a talent for messing up results,
perhaps from my years of torture testing my own software to try to eliminate
bugs. When I input information somewhat similar to my own situation, I
got a message of “unrealistic input.” In the cell for “Probability of a
Shortfall” I obtained “#NUM!”
However,
it does provide reasonable output in most scenarios. I already knew that
I have too much money accumulated for retirement, even though I stopped
contributing to retirement accounts in 1987.
This
software can be helpful both for clients who are already retired and for
those planning for retirement. If one is planning for retirement, a key
issue in calculating how much one needs to accumulate is the withdrawal
rate that is appropriate after retirement. If one purchases an annuity,
the nominal amount of income will be fixed, but there is an uncertain risk
of loss of purchasing power. If a retiree maintains a portfolio of stocks
and bonds, there is the dual challenge of finding the best portfolio allocation
and prudent withdrawal rate. In ancient times, one might decide to live
off dividends, but that is not very practical today. In any case, the issue
of how much a worker should save today for retirement depends on portfolio
allocation today, expected portfolio allocation or annuitization in retirement,
and the withdrawal rate in retirement. How
to Avoid Outliving Your Money is a very useful tool for such challenges.
The
program lists six asset classes, including: Canadian Equity, World Equity,
U.S. Equity, Canadian Bonds, and Canadian T-Bills. The developers promise
to provide a more specifically U.S. version next year, but the current
version would not be bad for U.S. classroom use right now. It is worthwhile
to examine the users manual available on the web site, as it is very informative.
The spreadsheet can be downloaded, but one must pay for the software to
obtain a keyword to unlock the file.
References
Cooley,
P. L., Hubbard, C. M. & Walz, D. T. (1999). Sustainable withdrawal
rates from your retirement portfolio. Financial Counseling and Planning,
10(1),
39-47.
Hanna,
S. (1998). Review of K. Ho, G. Perdue & C. Robinson (1998). Personal
Financial Planning. In Financial Counseling and Planning, 9
(2), 85-87.
Hanna,
S. & Chen, P. (1996). Efficient portfolios for saving for college,
Financial
Counseling and Planning, 7, 115-122.
ESPlanner:
2000, software, $64.95 for individual users, $495 for individual financial
planners. Publisher web site: www.esplanner.com
How to Avoid Outliving Your Money:2000,software, (US: $29.50,
Canadian: $39.50), Publisher web site: www.captus.com
Return to
journal home page