The document discusses key factors in retirement distribution planning, including longevity risks, market sequence risks, and risks related to maintaining sustainable withdrawal rates over long periods. It notes the importance of considering the dynamic interplay between a retiree's age, portfolio value and market performance over time when determining appropriate withdrawal strategies. A probabilistic approach focusing on the probability of portfolio failure at different withdrawal rates is presented as a way to assess risk in a dynamic framework.
1 of 41
More Related Content
Fpanc 2012
1. What are Key Factors in
Retirement Distribution Planning?
In a World with Longevity Risks, Failure Rates,
Market Sequence Risks, and Superannuation Risks.
Larry R. Frank, Sr. MBA, CFP®
BS cum laude Physics
Registered Investment Adviser (California)
Author, Wealth Odyssey
Presented to Financial Planning Association
of Northern California 14 September 2012
1 16:19
2. C-141 Cockpit: Levers and gauges
Everything is important,
otherwise it would not
be there.
What you pay attention to
depends on what you
are doing and the
situation (normal or
emergency).
Obviously yoke, throttles
and rudders, but those
need refined
measurements to be
useful.
How do you “fly” through three dimensions?
3. C-141 Cockpit: Levers and gauges
EMERGENCY PROCEDURES!
You cut consumption during an Fuel Consumption
engine emergency. (Withdrawal Rate)
But… not sensitive
enough … so need …
Another instrument?
Attitude indicator.
Which way is up?
Attitude and airspeed
interact with each
Airspeed indicator. other.
Too Slow – stall and crash.
Too Fast – run out of fuel and crash.
4. As Planners, what “levers and
gauges” do we have?
What can a planner really “control” for retiree (during
bad markets)?
Q1: When Markets go down: Does changing the portfolio
allocation work?
A1: No. Market timing doesn’t work (more on this
shortly).
Q2: When markets go down, does changing the
withdrawal dollar amount work?
A2: Yes. Most effective method to determine success or
failure.
Q3: Are buckets useful?
A3: A 3 bucket scenario is no more effective than a 2
bucket scenario, etc.. Sum of buckets equals a total allocation.
However, buckets are effective to address behavioral aspects
for retiree and to allow time to transition their spending.
5. Evolution of Research
Concepts today come from research summarized in this presentation.
Post Peer
Review.
Publication
Date TBA.
6. The Problem with 4%
A structural problem with pensions, immediate
annuities, and first generation "safe withdrawal rate" is
a disconnect between benefits paid (fixed or fixed with
COLAs) from the underlying asset values required to
support those promised benefits.
There are consequences to this disconnect, which we will
discuss shortly.
6 16:19
7. The Problem with 4%
You can’t tell one rate from another. Is a proper rate 4%?
Or 4.5? Or some other value?
In other words, withdrawal rates all look the same.
8. The Problem with 4%
“First Generation” research tries to shoehorn all
retirees into one “answer” (initial 4% and then add
inflation) … i.e., a static solution.
Ignores:
Time (which is dynamic)
Age changes with time (result: expected longevity changes)
Portfolio balances change with time (market effects)
Total Portfolio Value ($X) X Withdrawal Rate (WR%) = Annual Withdrawal ($Y)
9. Is withdrawal rate a good variable to use for
retirement decision rules?
WR% = $Y / $X
$Y =Dollars Withdrawn Poor Returns = WR%
$X = Portfolio Value 1 Market Returns
Good Returns = WR%
Withdrawal Rate WR%
Depends upon: Shorter Periods = WR%
2 Age
Longer Periods = WR%
1 Sequence Risk Paper 2010 (Probability of the Portfolio)
2 Longevity = Dynamic Distribution Period Paper 2012
(Probability of the Person)
These 2 factors will be a theme throughout.
Answer: No
10. Past perspective of Simulation
(Calculation) results
Graphs and figures create the illusion of knowing future outcomes.
We will see the range of possible futures, graphed later in the presentation,
that represent how the future may unfold with uncertainty (hint: we don’t know).
X%
First Generation Perspective: “Safe” Withdrawal Rate
applies through out the distribution period, i.e., a
“Calculate-and-forget” perspective.
Reality is not “static” … as the situation changes,
so do the assumptions and answers …
therefore solutions are dynamic
10 16:19
11. Second Generation: 3D & Dynamic perspective
of Simulation (Calculation) results
This is all a simulation (Calculation) tells us
(We do not know how the future will unfold). Retirees are transient –
This is the “solution” to the equation or simulation. they change position
w/in model as markets
change and as they age.
We need a model that
WR%
explains retirement for ALL
retirees (ages, allocations
and withdrawal rates)
when you realize that ALL
retirees co-exist at the same
time.
We will see what this
perspective looks like in 2
slides.
12. Resolving the 4% problem in a
dynamic world
What is the key “instrument,” or value, to focus on?
Probability of Failure (POF) rates
Defined here as the Percentage of simulations that FAIL to
reach the end of the simulation period
You will see later that simulation periods are dynamic to reflect the
reality of aging.
Inside a 3D Model - Dynamically (annually recalculated,
serially connected) Combines
Monte Carlo probability (probability of the portfolio) with
Longevity probability (probability of the person)
13. All Points on each Surface (“membrane” ) are the SAME
Three Dimensions Probability of Failure rate (POF); i.e.,
withdrawal rate changes, BUT POF the same=>need to know POF
in order to make sense of any withdrawal rate reference.
5% POF Surface 25% POF Landscape
11.00%-12.00% 11.00%-12.00%
12.00% 12.00%
11.00% 10.00%-11.00% 11.00% 10.00%-11.00%
10.00% 10.00%
9.00% 9.00%-10.00% 9.00%
8.00% 9.00%-10.00%
7.00% 8.00%
6.00% WR% 8.00%-9.00% 7.00% 8.00%-9.00%
5.00% 6.00% WR%
4.00% 7.00%-8.00% 5.00%
4.00% 7.00%-8.00%
3.00% 3.00%
2.00% 6.00%-7.00% 2.00% 6.00%-7.00%
1.00% 1.00%
0.00% 5.00%-6.00% 0.00% 5.00%-6.00%
0%
0%
30%
30%
60%
10
60%
90%
30 20 4.00%-5.00% 10
90%
40 40 30 20 4.00%-5.00%
3.00%-4.00% 3.00%-4.00%
Allocation (Equity%)Years Remaining Allocation (Equity%)Time Remaining
2.00%-3.00% 2.00%-3.00%
35% POF Landscape Not all retirees are the samePOF Landscape retirees co-
50% …AND all
exist (they transition through graphs based on
10.00%-11.00%
11.00%-12.00%
11.00%
10.00%
markets & time … they are transient).
9.00%-10.00% 12.00%
11.00% 10.00%-11.00%
9.00% 8.00%-9.00% 10.00%
8.00% 9.00% 9.00%-10.00%
7.00% 8.00%
WR%
6.00% 7.00%-8.00% 7.00%
5.00%
4.00%
- They are different ages (Time Axis). 8.00%-9.00% 6.00% WR%
3.00% 6.00%-7.00% 5.00%
2.00% - 7.00%-8.00%
They have different allocations (Allocation axis). 4.00%
1.00% 5.00%-6.00% 3.00%
0.00% 6.00%-7.00% 2.00%
- They spend more, or less, than they should (WR% axis)
4.00%-5.00% 1.00%
0.00%
0%
5.00%-6.00%
30%
0%
60%
15 10
30%
40 35 30 25 20 3.00%-4.00%
90%
60%
10
90%
40 30 20 4.00%-5.00%
Objective of the model is how to explain ALL
2.00%-3.00%
3.00%-4.00%
Allocation % Equity Time Remaining
retirees at ALL times. Time Remaining
1.00%-2.00% Allocation (Equity%)
2.00%-3.00%
13
14. THREE TIME/AGE CROSS SECTIONS THROUGH LANDSCAPES
Withdrawal rate goes up as age goes up.
However, Probability of Failure (POF)
pattern persists.
Withdrawal rate goes up as Portfolio
Values go down with declining markets
(sequence risk).
15. Withdrawal Rates (WR%) by age with all points calculated at 10% POF.
Looking at 3D graph along age axis so Allocations collapse into a single view.
Notice allocations below 60% equity are all very similar with “optimum” allocation
changing slowly as the retiree ages.
15 16:19
16. Application: Sequence Risk … Market goes down … what do you do?
Can pre-calculate portfolio values that represent possible future transient states
that would have increasing POF … AND the sustainable withdrawal Dollar amount
that would bring POF back to the original, lower, POF.
Life Expectancy Based on Outlive % and Retiring at
Age 70
54
Balanced allocation $3730 (Pre-tax; post fee) Joint Female Male
Outlive Life% 80% 85 74 70
$X * WR% = $Y Target POF 5% 31 20 16
WR% = $Y / $X Portfolio % Decline Probability of Failure (POF) @ WR%:
$X = $Y / WR% Required: WR%: POF Sustainable @ Lowest
Current LTP Value= $1,205,670 4.60% 5.00% POF @ Port. Value
$1,097,713 9.0% 5.00% 10.00% $ 3,381.00 Sustainable
Increasing POF $1,006,555 16.5% 5.40% 15.00% $ 3,097.25 Dist. $$ per
means moving up $937,196 22.3% 5.75% 20.00% $ 2,881.35 month.
through POF
landscapes or th
1st: Set POF 4 : Set a distribution
“membranes” 3rd: Determines Amount (back to 5%
2 nd: Determine WR%
Equivalent Portfolio POF) Emergency
Value Based on longevity percentile Procedure
for current age(s) = Distribution Period
Total Portfolio Value ($X) X Withdrawal Rate (WR%) = Annual Withdrawal ($Y)
17. There is a cost to high withdrawal
rates (high POF rates).
Cost comes from to many dollars withdrawn which
leads to a lower balance later.
Since future market returns are unknown, reaching for
return simply translates into higher volatility.
Research showed Lower Probability of Failure
(POF), i.e., Lower WR% => Higher Lifetime Cash
flows and Higher Terminal Values
Question in the past has been: How do you know when a
withdrawal rate is too high?
Answer: When POF approaches 30% (JFP 2011 paper
using control data methodology).
17 16:19
18. WHAT IS ANOTHER KEY FACTOR TO DETERMINE THE WITHDRAWAL RATE?
Answer: The Distribution Period (DP)
What determines the length of the Distribution Period?
Answer: Longevity.
However, can not determine longevity just once … because it changes slightly
every year as the retiree(s) age.
Therefore, this is a dynamic process you revisit each year as they age
(example coming).
18 16:19
19. Probability of Longevity By surviving to an older age it becomes MORE likely
You will survive to an even older age.
Period Life Chart
Probability of survival past a given ”mortality age” from a base “living age”
100%
90%
80%
70% "Target" or
mortality age
60% B
75
50%
80
40% A
A 85
30%
90
20% 95
10% 100
0%
50 55 60 65 70 75 80 85 90 95 100
A Typical retirement ages
Living age
Source: United States Life Tables 2003. National Vital Statistics Reports, Vol. 54, Number 14, revised
Common 28 2007, Table 10: All races, 2003.
March
Perception of Probability of the Person
Age anchors
20. Longevity Dynamics
Death is always before expected longevity for the
current age; in other words current age, and expected
longevity age, BOTH change each year.
Withdrawal Rate (WR%)
WR%
|
Today’s Age End Age
Withdrawal Period (WP)
How much time remains?
20
21. Putting it all together
Let’s put it all together.
21 16:19
24. RETIREE CASH FLOWS AGES 60, 70, 80 & 90 (WITH SUPERANNUATION ADJUSTMENT (1/N) )
RESULT OF START WITH EXPECTED LONGEVITY PERCENTILE, DYNAMICALLY ADJUSTING OVER TIME
TO LOWER LONGEVITY PERCENTILE (EFFECT: DYNAMICALLY EXTENDS THE DISTRIBUTION PERIOD)
At each age: Withdrawal rate is
the same; Different $$ amount
which depends on good or bad We would like
market sequence. “good”
markets.
? We need to plan
2000’s ? for “poor”
markets.
Serially connected, Annually Adjusted, Withdrawals. All Points calculated at 10% POF; 60 % Equity
25. Higher withdrawals early
(50th (30%))(Light Red line).
Result in lower withdrawals later …
Compare to 50th (10%) (Dark green
line).
Conclusion:
Trying to squeeze out more ….
results in getting less over their
lifetime.
25
26. Second Generation: 3D & Dynamic perspective
of Simulation (Calculation) results
This is all a simulation (Calculation) tells us
(We do not know how the future will unfold)
WR%
We don’t need complex simulations
to project the future. We just need to
know the present facts and use the
concepts from research to pre-calculate
acceptable parameters based on facts
we know today.
X%
“Current” Withdrawal Rate (WR%)
applies only for the moment the simulation (or calculation) is run
26 16:19
27. Key basic takeaway
Life Expectancy Based on Outlive % and Retiring at
Age 70
54
Balanced allocation $3730 (Pre-tax; post fee) Joint Female Male
Outlive Life% 80% 85 74 70
$X * WR% = $Y Target POF 5% 31 20 16
WR% = $Y / $X Portfolio % Decline Probability of Failure (POF) @ WR%:
$X = $Y / WR% Required: WR%: POF Sustainable @ Lowest
Current LTP Value= $1,205,670 4.60% 5.00% POF @ Port. Value
$1,097,713 9.0% 5.00% 10.00% $ 3,381.00 Sustainable
$1,006,555 16.5% 5.40% 15.00% $ 3,097.25 Dist. $$ per
$937,196 22.3% 5.75% 20.00% $ 2,881.35 month.
Emergency
1st: Set POF Procedures!
2nd: Determine WR%
Based on longevity percentile
for current age(s) = Distribution Period
Total Portfolio Value ($X) X Withdrawal Rate (WR%) = Annual Withdrawal ($Y)
29. 4 Levers: A summary – popping open withdrawal
rates to look inside them
Sequence Risk (unknown market returns) Lever*
When POF approaches 30%
Distribution Period Lever (tilt towards consumption or bequest
goals)*
Use table Longevity Percentiles of Period Life Tables
Superannuation Lever (Risk of very old age) (2013 Paper)
Transition those in mid-retirement (mid to late 70’s) to expect older
ages; thus a continued extension of distribution periods
Portfolio Volatility Lever
Allocation (least effective)
*Withdrawal rate (WR%) most sensitive to these two levers
(purpose of monitoring WR% - or more specifically POF: Higher
cash flow early results in lower cash flow later because portfolio
balances are consumed & can not assume returns will over come
that consumption).
29 16:19
30. So what control really matters?
Remember earlier we discussed combining TWO sets of probabilities that interact:
Probability of Failure – the percent of simulations that fail to reach the end of the
simulation period given present withdrawal rate.
(Probability of the Portfolio … or sequence risk)
Simulation period is a function of current age & current age’s longevity percentile.
(Probability of the Person)
Navigation
(Are you on
track towards
Attitude Indicator destination?)
Airspeed Consumption
(by itself not
an effective
measure)
31. The solution to expand beyond the
static 4% perspective
Incorporate a 3-D perspective
Second Generation is Dynamic (rather than static)
A sustainable methodology needs to keep benefits
connected to supporting asset values year by year
throughout the entire distribution period.
32. Summary: Blend of Expert Team
and knowing what levers to move
Annual review and
update based on facts
as they are then known.
Emergency procedures
well understood by
you and retiree.
Coordination of team
of other experts so
retiree successfully “flies
through life.”
33. The
End
… RESEARCH STILL DIGGING DEEPER INTO STILL
UNANSWERED QUESTIONS …
PS. Presentation posted 14 September 2012 may be found
at blog.betterfinancialeducation.com or http://www.slideshare.net
33 16:19
34. Disclosures
There can be no assurance that the financial concepts and
strategies presented in this material will be successful.
Investments are subject to market fluctuation, market
risk, and loss of principal. Past performance is not a
guarantee of future success. Every investment strategy has
the potential for profit or loss.
For allocations, please refer to the research material
referenced at the slides in Background info.
Investors cannot invest directly in an index.
Therefore, passive indexed approaches have been
developed. The performance of any index is not indicative
of the performance of any investment and does not take
into account the effects of inflation and the fees and
expenses associated with investing.
36. We have gone from “Classical” withdrawal rate thought (4%)
Newtonian Physics:
Deterministic
determination of
future positions.
To a deeper insight into what goes on “inside” a withdrawal
rate.
Quantum Physics:
Probability of future, but
not which future actually
ensues.
37. Decision Rules
Total Portfolio Value ($X) X Withdrawal Rate (WR%) = Annual Withdrawal ($Y)
$X = $Y
Guyton WR%
20% increase in WR% (Capital Preservation Rule)*
20% decrease in WR% (Prosperity Rule)*
Measured from when? (High? Middle? Low?)
Do the rules apply:
over all periods?
to all asset allocations?
Are there more
consistent rules based
on dynamic principles?
*JFP March 2006
38. Timeline and POF
•POF is how one evaluates the future possibilities
PAST PRESENT FUTURE
100%
Probability of Portfolio Failure Based on Remaining
90%
80%
70%
60%
Target Period
50%
40%
30%
20%
10%
Today
0%
0 5 10 15 20 25 30
Elapsed Time (Years)
Black to cover the years up Runs That Passed Runs That Failed
3 4
5 6
7 8
9 10
11 12
13 14
15 16
17 18
HISTORY (KNOWN) TOMORROW’S EVENTS(UNKNOWN) 19
21
23
20
22
24
25 26
27 28
•Simulations may be run for any POF (illustrated at 40%) 29
31
33
30
32
34
35 36
•Let us see how different POF’s graph in comparison 37
39
41
38
40
42
43 44
45 46
47 48
49 50
38
40. A Dynamic and Adaptive Approach to Distribution Planning and
Monitoring. April 2009, by David M. Blanchett and Larry R Frank
Sr.
• Demonstrates how dynamically changing the withdrawal
dollar amount leads to higher withdrawal rates. Questions:
• When to make the withdrawal dollar amount change?
• Does changing allocation make a difference?
The Dynamic Implications of Sequence Risk on a Distribution
Portfolio. June 2010, by Larry R Frank Sr and David M. Blanchett.
• Demonstrates that sequence risk does NOT go away with time.
Question:
• What should decision rules be based on if sequence risk is ever-
present?
40
41. Working papers with research data may be found on Social Science Research Network
(SSRN.com) Quicksearch “Larry R Frank”
Probability-of-Failure-Based Decision Rules to Manage
Sequence Risk in Retirement. November 2011, by Larry R
Frank Sr, John B Mitchell, and David M. Blanchett.
An Age-Based, Three-Dimensional Distribution Model
Incorporating Sequence and Longevity Risks. March
2012, by Larry R Frank Sr, John B Mitchell, and David M.
Blanchett.
Academy of Financial Servicers October 1-2, 2012 at The Menger
Hotel, San Antonio, TX.,
Transition from Young, through Very Elderly, Retirement
Distributions within the Age Based, 3D Universal
Distribution Model
by Larry R Frank Sr, John B Mitchell, and David M. Blanchett.
Superannuation risk.
41
Editor's Notes
Lots of controls, but what really mattersShare trainingEt al
Lots of controls, but what really mattersShare trainingEt al
Discuss research at high level
I’m my clients co-pilot and navigator helping them achieve a successful retirement.