30 years not long enough to calculate SWR?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
Raybo
Posts: 1861
Joined: Tue Feb 20, 2007 11:02 am
Location: San Francisco
Contact:

30 years not long enough to calculate SWR?

Post by Raybo » Sat Feb 16, 2019 6:11 pm

I started reading Larry Swedroe's new book (co-authored with Kevin Grogan) Your Complete Guide to a Successful and Secure Retirement. In it, he says
Historically, a safe harbor withdrawal rate for a 65 year-old couple has only been about 4% of the portfolio's initial value [adjusted each year for inflation]. As we have discussed, given today's lower bond yields, lower expected stock returns, greater longevity, and greater need to save for medical expenses, that need may be too aggressive for many.
I suspect that there have been years in the past where the combination of lower bond yields and lower expected stock returns have existed, so that contingency should be included in the Trinity Study (I don't have the data, so can't check this on my own).

This would mean the main point Larry and Kevin are making here is that an SWR of 4% may be "too aggressive," because retirement may need to last more than 30 years (the book states that there is a 25% chance that one of a couple might live to 97) and expenses as one ages go up due to increasing medical expenses.

So, the issue with the Trinity Study isn't the 4% rate, but the fact that it only goes out 30 years and only applies general inflation to expenses, not some higher medical expense inflation rate.

There's some non-zero chance that I or my wife might live to be 120. There are real (life style) costs to limiting my withdrawals to some arbitrary percentage, say 3%, without some basis for doing so.

My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
No matter how long the hill, if you keep pedaling you'll eventually get up to the top.

User avatar
FiveK
Posts: 8000
Joined: Sun Mar 16, 2014 2:43 pm

Re: 30 years not long enough to calculate SWR?

Post by FiveK » Sat Feb 16, 2019 6:20 pm

Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
There is no deterministic answer, because the future is unknown and unknowable.

See (among other threads and wiki articles) Mr. Money Mustache, SWR, and equity allocation - Bogleheads.org for much discussion about probability, likelihood, risk assessment, etc., as it pertains to this question.

dbr
Posts: 31155
Joined: Sun Mar 04, 2007 9:50 am

Re: 30 years not long enough to calculate SWR?

Post by dbr » Sat Feb 16, 2019 6:29 pm

You can test the sensitivity to input assumptions. In short, run some models for longer durations and see what the effect is. That doesn't tell you what length of time to plan for but it does tell you how much it matters what length of time to plan for. The actual effect in most models is that the SWR decreases at a slowing rate for longer times and eventually reaches an asymptote which might be called the perpetual withdrawal rate. It could be there is a small effect on the asset allocation one might choose.

Naturally the further into the future one wants to project the more uncertainty must develop, not just in investment projections but in secular conditions and personal circumstances.

GrowthSeeker
Posts: 785
Joined: Tue May 15, 2018 10:14 pm

Re: 30 years not long enough to calculate SWR?

Post by GrowthSeeker » Sat Feb 16, 2019 6:50 pm

There is also this website, Early Retirement Now, where the author did an extensive study of SWR based on monthly past market data and longer periods of retirement up to 60 years.
Just because you're paranoid doesn't mean they're NOT out to get you.

AlohaJoe
Posts: 5060
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: 30 years not long enough to calculate SWR?

Post by AlohaJoe » Sat Feb 16, 2019 7:32 pm

Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You should use the joint mortality of your and your partner, taking into account your actual ages. Then take the 80th percentile. Use a website like longevityillustrator from the Society of Actuaries.

You don't need to factor in higher medical expenses. Other expenses go down more than medical expenses go up.

User avatar
SevenBridgesRoad
Posts: 582
Joined: Sat Jul 07, 2018 12:14 am
Location: Within Pacific Ring of Fire

Re: 30 years not long enough to calculate SWR?

Post by SevenBridgesRoad » Sat Feb 16, 2019 7:54 pm

FiveK wrote:
Sat Feb 16, 2019 6:20 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
There is no deterministic answer, because the future is unknown and unknowable.

See (among other threads and wiki articles) Mr. Money Mustache, SWR, and equity allocation - Bogleheads.org for much discussion about probability, likelihood, risk assessment, etc., as it pertains to this question.
And, as if all of that isn't enough to confuse your planning, toss in the 'timeline to the singularity'. There is a non-zero chance we will no longer need health care at all. The future might be wildly unimaginable.

https://futurism.com/kurzweil-claims-th ... en-by-2045
Retired 2018 age 61 | "Not using an alarm is one of the great glories of my life." Robert Greene

OffGridder
Posts: 51
Joined: Thu Jul 23, 2015 8:03 am
Location: Eastern WA.

Re: 30 years not long enough to calculate SWR?

Post by OffGridder » Sat Feb 16, 2019 7:56 pm

AlohaJoe wrote:
Sat Feb 16, 2019 7:32 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You should use the joint mortality of your and your partner, taking into account your actual ages. Then take the 80th percentile. Use a website like longevityillustrator from the Society of Actuaries.

You don't need to factor in higher medical expenses. Other expenses go down more than medical expenses go up.
I do not believe it is reasonable to plan or assume the reduction or elimination of discretionary spending in old age will offset the cost of Long Term Care if needed.
"Goodness is the only investment that never fails." | H.D. Thoreau

2015
Posts: 2906
Joined: Mon Feb 10, 2014 2:32 pm

Re: 30 years not long enough to calculate SWR?

Post by 2015 » Sat Feb 16, 2019 10:32 pm

FiveK wrote:
Sat Feb 16, 2019 6:20 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
There is no deterministic answer, because the future is unknown and unknowable.

See (among other threads and wiki articles) Mr. Money Mustache, SWR, and equity allocation - Bogleheads.org for much discussion about probability, likelihood, risk assessment, etc., as it pertains to this question.
+1

AlohaJoe
Posts: 5060
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: 30 years not long enough to calculate SWR?

Post by AlohaJoe » Sun Feb 17, 2019 3:48 am

OffGridder wrote:
Sat Feb 16, 2019 7:56 pm
AlohaJoe wrote:
Sat Feb 16, 2019 7:32 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You should use the joint mortality of your and your partner, taking into account your actual ages. Then take the 80th percentile. Use a website like longevityillustrator from the Society of Actuaries.

You don't need to factor in higher medical expenses. Other expenses go down more than medical expenses go up.
I do not believe it is reasonable to plan or assume the reduction or elimination of discretionary spending in old age will offset the cost of Long Term Care if needed.
So you don't have a constructive answer for the OP's question?

Tell him how to factor it in, if you don't like my answer. Just telling him you disagree with my answer doesn't actually help him in any way.

msk
Posts: 1439
Joined: Mon Aug 15, 2016 10:40 am

Re: 30 years not long enough to calculate SWR?

Post by msk » Sun Feb 17, 2019 4:21 am

It's always been silly to adjust withdrawals to inflation without incorporating market performance. The 4% rule is useful as a trigger to identify FIRE. One way to ensure that my portfolio will last forever is to limit the percentage of portfolio I withdraw in any year. My rule is never to withdraw more than 5% of portfolio balance, ever. The portfolio will last forever, of course, and if the stock market behavior resembles the history since 1871, then a WR of 5% will deliver a spend that keeps up with inflation on average despite yoyoing up and down with the market, and leave a portfolio balance that keeps up with inflation forever. Too much yoyoing with the market? Use 3-year rolling averages, like some university endowment funds. Still terrified because one has saved barely enough? Keep 3 years' spend in bonds and simply roll it over until you experience a huge stock market fall.

By the way, exactly what will happen if I am broke and old and cannot afford LTC or whatever. Die quickly? Is that so bad?

Leesbro63
Posts: 6081
Joined: Mon Nov 08, 2010 4:36 pm

Re: 30 years not long enough to calculate SWR?

Post by Leesbro63 » Sun Feb 17, 2019 6:48 am

msk wrote:
Sun Feb 17, 2019 4:21 am
It's always been silly to adjust withdrawals to inflation without incorporating market performance. The 4% rule is useful as a trigger to identify FIRE. One way to ensure that my portfolio will last forever is to limit the percentage of portfolio I withdraw in any year. My rule is never to withdraw more than 5% of portfolio balance, ever. The portfolio will last forever, of course, and if the stock market behavior resembles the history since 1871, then a WR of 5% will deliver a spend that keeps up with inflation on average despite yoyoing up and down with the market, and leave a portfolio balance that keeps up with inflation forever. Too much yoyoing with the market? Use 3-year rolling averages, like some university endowment funds. Still terrified because one has saved barely enough? Keep 3 years' spend in bonds and simply roll it over until you experience a huge stock market fall.

By the way, exactly what will happen if I am broke and old and cannot afford LTC or whatever. Die quickly? Is that so bad?
Actually if you’re old and broke and need LTC, the state will provide it. The problem is if you’re old and broke but healthy.

Dandy
Posts: 6050
Joined: Sun Apr 25, 2010 7:42 pm

Re: 30 years not long enough to calculate SWR?

Post by Dandy » Sun Feb 17, 2019 7:35 am

I think there is no set it and forget it withdrawal strategy as much as we all would like that. Too many variables with a several decades retirement given changes in investments, interest rates, inflation, taxes, medical expenses, long term care possibilities, etc. So, pick a reasonable approach to start with e.g. 4% maybe a bit less and keep an eye on your portfolio, expenses, health etc. and if expenses are starting to accelerate cut back on them/reduce withdrawals a bit. If portfolio is growing and expenses are under control withdraw a bit more. In other words manage your money as opposed to "outsourcing" it to a formula. Also, if your pension/SS income is rather low compared to your normal expenses consider an immediate annuity when you are 80 or so.

Having said that a 3 or 4% withdrawal rate is most likely going to last 30 years or so most of the time. So, it isn't a bad starting point - but that is really all that it is since the future is unknown. So it is more of set it and keep an eye on it.

Stormbringer
Posts: 889
Joined: Sun Jun 14, 2015 7:07 am

Re: 30 years not long enough to calculate SWR?

Post by Stormbringer » Sun Feb 17, 2019 8:46 am

Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You could buy an inflation-adjusted SPIA that pays 4% for life. Unless you are determined to leave a bundle for your kids, this seems like a better option to me, at least for basic retirement expenses.
"Compound interest is the most powerful force in the universe." - Albert Einstein

Leesbro63
Posts: 6081
Joined: Mon Nov 08, 2010 4:36 pm

Re: 30 years not long enough to calculate SWR?

Post by Leesbro63 » Sun Feb 17, 2019 9:06 am

Stormbringer wrote:
Sun Feb 17, 2019 8:46 am
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You could buy an inflation-adjusted SPIA that pays 4% for life. Unless you are determined to leave a bundle for your kids, this seems like a better option to me, at least for basic retirement expenses.
These are hugely expensive.

User avatar
Sandtrap
Posts: 8627
Joined: Sat Nov 26, 2016 6:32 pm
Location: Hawaii No Ka Oi , N. Arizona

Re: 30 years not long enough to calculate SWR?

Post by Sandtrap » Sun Feb 17, 2019 9:14 am

AlohaJoe wrote:
Sat Feb 16, 2019 7:32 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You should use the joint mortality of your and your partner, taking into account your actual ages. Then take the 80th percentile. Use a website like longevityillustrator from the Society of Actuaries.

You don't need to factor in higher medical expenses.Other expenses go down more than medical expenses go up.
(with my luck, I've landed outside of these "percentile numbers" :shock: )

Good points as always.
Additionally. . . .
Perhaps, just perhaps, just as point mortality is taken into account, there's also family medical concerns/lineage. IE: cancer, long term terminal illness and care vs those that have a tendency to "go quick", disability, Parkinsons, etc, etc.
But, that is hard to predict, if predict quantifiably at all, as OP suggests. All of us know instances where medical issues quickly or gradually eroded retirement savings far quicker than anything else. Real life examples of 30X destroyed by financial and personal Black Swans.
I've seen this often in my extended family (generations) and many with substantial wealth so it's a concern to me as well.
(perhaps not per forum guidelines but, tangentially the evolving nature of Health Care and Insurance is also a concern)

As others have mentioned, the longer the prediction, the more variables, the less able to predict.

How can future financial and personal and/or health cost "black swans" be calculated into things?

Actionably: per OP: All you can do is try 3%, evaluate, adjust . . try 3.5%, evaluate, adjust . . .repeat. Variable Withdrawal Rate per your evolving lifestyle and financial and personal health circumstance.
Wiki Bogleheads Wiki: Everything You Need to Know

Stormbringer
Posts: 889
Joined: Sun Jun 14, 2015 7:07 am

Re: 30 years not long enough to calculate SWR?

Post by Stormbringer » Sun Feb 17, 2019 9:54 am

Leesbro63 wrote:
Sun Feb 17, 2019 9:06 am
Stormbringer wrote:
Sun Feb 17, 2019 8:46 am
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You could buy an inflation-adjusted SPIA that pays 4% for life. Unless you are determined to leave a bundle for your kids, this seems like a better option to me, at least for basic retirement expenses.
These are hugely expensive.
$100K in a portfolio @ 4% SWR
$100K in an CPI-adjusted SPIA @ 4% payout ratio

I'm not seeing the huge cost difference there.
"Compound interest is the most powerful force in the universe." - Albert Einstein

tibbitts
Posts: 9397
Joined: Tue Feb 27, 2007 6:50 pm

Re: 30 years not long enough to calculate SWR?

Post by tibbitts » Sun Feb 17, 2019 10:02 am

Leesbro63 wrote:
Sun Feb 17, 2019 9:06 am
Stormbringer wrote:
Sun Feb 17, 2019 8:46 am
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You could buy an inflation-adjusted SPIA that pays 4% for life. Unless you are determined to leave a bundle for your kids, this seems like a better option to me, at least for basic retirement expenses.
These are hugely expensive.
I'm not sure if they're expensive but I'm not sure all of them are actually inflation-protected. I haven't read the fine print but do all of them pay out at unlimited levels of inflation?

dbr
Posts: 31155
Joined: Sun Mar 04, 2007 9:50 am

Re: 30 years not long enough to calculate SWR?

Post by dbr » Sun Feb 17, 2019 10:39 am

tibbitts wrote:
Sun Feb 17, 2019 10:02 am
I'm not sure if they're expensive but I'm not sure all of them are actually inflation-protected. I haven't read the fine print but do all of them pay out at unlimited levels of inflation?
[/quote]

There are annuities that have a fixed rather than inflation indexed COLA. Such an annuity is not actually inflation protected.

User avatar
Watty
Posts: 18012
Joined: Wed Oct 10, 2007 3:55 pm

Re: 30 years not long enough to calculate SWR?

Post by Watty » Sun Feb 17, 2019 10:52 am

dbr wrote:
Sat Feb 16, 2019 6:29 pm
You can test the sensitivity to input assumptions. In short, run some models for longer durations and see what the effect is. That doesn't tell you what length of time to plan for but it does tell you how much it matters what length of time to plan for. The actual effect in most models is that the SWR decreases at a slowing rate for longer times and eventually reaches an asymptote which might be called the perpetual withdrawal rate. It could be there is a small effect on the asset allocation one might choose.

Naturally the further into the future one wants to project the more uncertainty must develop, not just in investment projections but in secular conditions and personal circumstances.
+1

It has been decades since I had the college classes that covered sensitivity analysis but what I did when I was deciding if I had enough to retire was to use web sites like Firecalc to model my spending using a 4% withdrawal rate until I was 95.

The brute force "sensitivity analysis" that I did was to increase and decrease various combinations of withdrawal rates and ages and I entered those into a spreadsheet which had the age and rows with the corresponding success rate.

https://www.firecalc.com/

For example a 4% withdrawal rate for 30 years has a 95% success rate using the Firecalc defaults.

For 35 years it is 93%, 40 years 85%

With a 4.1% withdraw rate for 30 years it is 93%, 4.3% is 91.5%, etc

I don't have the spreadsheet that I did anymore but I ended up with a grid that showed the various success rates with the combinations of ages from 90 to 105 and withdrawal rates from about 3.5% to 5%.

What I was looking for was not so much the specific success rate for any combination but to see if there was a sharp drop off where a small change caused big reduction in the success rate.

There was not any sharp drop off so that helped me feel better about being able to be financially OK if I live to be 105.

The risk of running out of money does increase some each year I live past 95 but you have to multiply that by the odds of you living an extra year so the combined odds quickly get to be around 99% that you don't live long enough to run out of money.

Whenever you look at safe withdrawal rate you also need to keep in mind that the Trinity study was just an academic study and the authors have even said that it was not meant to be used a real world spending model. Your spending will not be the same each year and unless you are planning a barebones retirement you will also be able to reduce your spending if your portfolio is being depleted faster than expected. If you had to make a 5% reduction in spending that would take you from a 4% withdrawal rate to 3.8%. That would make a huge difference in projected success rates but only a minor difference in most peoples lifestyles.
Last edited by Watty on Sun Feb 17, 2019 10:55 am, edited 2 times in total.

Leesbro63
Posts: 6081
Joined: Mon Nov 08, 2010 4:36 pm

Re: 30 years not long enough to calculate SWR?

Post by Leesbro63 » Sun Feb 17, 2019 10:54 am

Stormbringer wrote:
Sun Feb 17, 2019 9:54 am
Leesbro63 wrote:
Sun Feb 17, 2019 9:06 am
Stormbringer wrote:
Sun Feb 17, 2019 8:46 am
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You could buy an inflation-adjusted SPIA that pays 4% for life. Unless you are determined to leave a bundle for your kids, this seems like a better option to me, at least for basic retirement expenses.
These are hugely expensive.
$100K in a portfolio @ 4% SWR
$100K in an CPI-adjusted SPIA @ 4% payout ratio

I'm not seeing the huge cost difference there.
The cost difference is that for the SWR thing, most likely there will be a huge pot for heirs and maybe for "beyond inflation" annual spending increases later in the owner's life. And you'll have the actual cash for a change in plans etc. Once you buy that annuity, it's gone and you have a fixed plan forever. You trade A LOT of future possibility for a little certainty.
Last edited by Leesbro63 on Sun Feb 17, 2019 12:07 pm, edited 1 time in total.

User avatar
willthrill81
Posts: 15054
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: 30 years not long enough to calculate SWR?

Post by willthrill81 » Sun Feb 17, 2019 11:01 am

Raybo wrote:
Sat Feb 16, 2019 6:11 pm
This would mean the main point Larry and Kevin are making here is that an SWR of 4% may be "too aggressive," because retirement may need to last more than 30 years (the book states that there is a 25% chance that one of a couple might live to 97) and expenses as one ages go up due to increasing medical expenses.
I'm not sure where Larry got the data from, but it doesn't jive with the Social Security Administration's life expectancy data. Kitces has an Excel spreadsheet based on the SSA data, and it indicates that for an opposite gender couple both aged 65, there is only a 16% probability that at least one of them will survive to age 95.

If you carry that forward to age 97, there is only a 9% chance of either of them still being alive, not 25%.

Granted, the typical Boglehead probably has a longer life expectancy than the general populace, which is what the SSA data are based on, but I'm not aware of a more comprehensive data set available to the public.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

User avatar
Tyler9000
Posts: 476
Joined: Fri Aug 21, 2015 11:57 am

Re: 30 years not long enough to calculate SWR?

Post by Tyler9000 » Sun Feb 17, 2019 11:21 am

Raybo wrote:
Sat Feb 16, 2019 6:11 pm
There's some non-zero chance that I or my wife might live to be 120. There are real (life style) costs to limiting my withdrawals to some arbitrary percentage, say 3%, without some basis for doing so.

My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint?
You're correct to note that safe withdrawal rates are calculated over a specific length of time, which makes them less than ideal for very long retirements where the endpoint is particularly uncertain. For retirement periods more than 30 years, I personally prefer to study the perpetual withdrawal rate that is designed to preserve your inflation-adjusted principal even in the worst case. Stick to that, and your portfolio will theoretically last forever.

For more info, you may find this useful: Perpetual Withdrawal Rates Are The Runway To A Long Retirement

OffGridder
Posts: 51
Joined: Thu Jul 23, 2015 8:03 am
Location: Eastern WA.

Re: 30 years not long enough to calculate SWR?

Post by OffGridder » Sun Feb 17, 2019 11:55 am

AlohaJoe wrote:
Sun Feb 17, 2019 3:48 am
OffGridder wrote:
Sat Feb 16, 2019 7:56 pm
AlohaJoe wrote:
Sat Feb 16, 2019 7:32 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You should use the joint mortality of your and your partner, taking into account your actual ages. Then take the 80th percentile. Use a website like longevityillustrator from the Society of Actuaries.

You don't need to factor in higher medical expenses. Other expenses go down more than medical expenses go up.
I do not believe it is reasonable to plan or assume the reduction or elimination of discretionary spending in old age will offset the cost of Long Term Care if needed.
So you don't have a constructive answer for the OP's question?

Tell him how to factor it in, if you don't like my answer. Just telling him you disagree with my answer doesn't actually help him in any way.
If the OP is budgeting $75K - $100K per year for discretionary spending in today's dollars, than the OP could reasonably plan to use that amount adjusted for inflation to offset the cost of LTC if needed in the latter years of retirement.
Last edited by OffGridder on Sun Feb 17, 2019 12:38 pm, edited 1 time in total.
"Goodness is the only investment that never fails." | H.D. Thoreau

User avatar
willthrill81
Posts: 15054
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: 30 years not long enough to calculate SWR?

Post by willthrill81 » Sun Feb 17, 2019 12:25 pm

Raybo wrote:
Sat Feb 16, 2019 6:11 pm
There's some non-zero chance that I or my wife might live to be 120.
Be careful with that logic. It leads down a rabbit hole with no exit. If history is any indicator of the future, there is a non-zero chance of all sorts of events occurring which would completely derail any type of retirement withdrawal plan we can form.

All roads carry risk.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

User avatar
SevenBridgesRoad
Posts: 582
Joined: Sat Jul 07, 2018 12:14 am
Location: Within Pacific Ring of Fire

Re: 30 years not long enough to calculate SWR?

Post by SevenBridgesRoad » Mon Feb 18, 2019 11:40 pm

willthrill81 wrote:
Sun Feb 17, 2019 12:25 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
There's some non-zero chance that I or my wife might live to be 120.
Be careful with that logic. It leads down a rabbit hole with no exit. If history is any indicator of the future, there is a non-zero chance of all sorts of events occurring which would completely derail any type of retirement withdrawal plan we can form.

All roads carry risk.
I'll offer a companion opinion: Not every potential future is a derail-er. There's better than a non-zero chance many aspects of the future will be better and cheaper. Some roads lead to dramatic improvements at lower costs. For example, we are not terribly far from curing diseases at the genetic level. With progress from the $5 billion dollar Human Genome Project just 20 years ago to today's desktop gene sequencers to incredible emerging therapies, we are on an exponential curve to improve the human condition with respect to disease.

No Pollyanna here. I agree, one must reasonably cover future risks best we can. But, don't confuse that with the likelihood of a worse, more expensive future.

There are infinite non-zeros, but not all of them lead to disaster.
Retired 2018 age 61 | "Not using an alarm is one of the great glories of my life." Robert Greene

User avatar
willthrill81
Posts: 15054
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: 30 years not long enough to calculate SWR?

Post by willthrill81 » Tue Feb 19, 2019 12:37 am

SevenBridgesRoad wrote:
Mon Feb 18, 2019 11:40 pm
willthrill81 wrote:
Sun Feb 17, 2019 12:25 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
There's some non-zero chance that I or my wife might live to be 120.
Be careful with that logic. It leads down a rabbit hole with no exit. If history is any indicator of the future, there is a non-zero chance of all sorts of events occurring which would completely derail any type of retirement withdrawal plan we can form.

All roads carry risk.
I'll offer a companion opinion: Not every potential future is a derail-er. There's better than a non-zero chance many aspects of the future will be better and cheaper. Some roads lead to dramatic improvements at lower costs. For example, we are not terribly far from curing diseases at the genetic level. With progress from the $5 billion dollar Human Genome Project just 20 years ago to today's desktop gene sequencers to incredible emerging therapies, we are on an exponential curve to improve the human condition with respect to disease.

No Pollyanna here. I agree, one must reasonably cover future risks best we can. But, don't confuse that with the likelihood of a worse, more expensive future.

There are infinite non-zeros, but not all of them lead to disaster.
That was the basic point I was trying to convey. Thank you for clarifying.

We don't have to think too long and hard to find flaws in literally any withdrawal strategy. What if you die young? Then the SPIA turned out to be a poor choice. What if you live to be 120? Then your VPW plan left you high and dry (unless you bought a SPIA). And so on.

We must balance the various risks we face according to our own best judgment.
“It's a dangerous business, Frodo, going out your door. You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.” J.R.R. Tolkien,The Lord of the Rings

Rob1
Posts: 172
Joined: Thu May 04, 2017 2:57 pm

Re: 30 years not long enough to calculate SWR?

Post by Rob1 » Tue Feb 19, 2019 2:32 am

OffGridder wrote:
Sat Feb 16, 2019 7:56 pm
AlohaJoe wrote:
Sat Feb 16, 2019 7:32 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
My question is how one might actually implement this information for planning purposes? How many years should I use for an endpoint? How to factor in higher medical expenses?
You should use the joint mortality of your and your partner, taking into account your actual ages. Then take the 80th percentile. Use a website like longevityillustrator from the Society of Actuaries.

You don't need to factor in higher medical expenses. Other expenses go down more than medical expenses go up.
I do not believe it is reasonable to plan or assume the reduction or elimination of discretionary spending in old age will offset the cost of Long Term Care if needed.
For what it is worth, I have a parent in assisted living. So far, the total annual cost (facilty, clothing...everything) is about the same as when my parent was living at home. Addionally, there are proceeds from selling a home that will help fund care. However, I do expect costs to increase significantly yearly, and to surpass what costs would have been if living at home.

User avatar
Watty
Posts: 18012
Joined: Wed Oct 10, 2007 3:55 pm

Re: 30 years not long enough to calculate SWR?

Post by Watty » Tue Feb 19, 2019 6:47 am

willthrill81 wrote:
Sun Feb 17, 2019 12:25 pm
Raybo wrote:
Sat Feb 16, 2019 6:11 pm
There's some non-zero chance that I or my wife might live to be 120.
Be careful with that logic. It leads down a rabbit hole with no exit. If history is any indicator of the future, there is a non-zero chance of all sorts of events occurring which would completely derail any type of retirement withdrawal plan we can form.

All roads carry risk.
It is also good to keep in mind that a 4% safe withdrawal rate is very conservative so it has a build in margin of safety that is a lot larger than the sum of the possible non-zero event chances like extended lifespans.

Using the defaults at Firecalc the average balance after 30 years for someone starting out with a million dollars is $1.8 million dollars. Using a 55 year withdrawal period would get a 65 year old to the age of 120 and it would still have a 79.6% successes rate according to firecalc.

This makes sense because a large percentage of people that retire at 65 will get lucky and have good investing returns for the first ten year of their retirement and then be making withdrawals at less than a perpetual safe withdrawal rate.

Post Reply