Variable Percentage Withdrawal (VPW)

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
travelintime
Posts: 8
Joined: Sun Mar 15, 2020 7:04 pm

Re: Variable Percentage Withdrawal (VPW)

Post by travelintime »

Thanks for your response
Lastrun
Posts: 345
Joined: Wed May 03, 2017 6:46 pm

Re: Variable Percentage Withdrawal (VPW)

Post by Lastrun »

First Longinvest I want to thank you for all your hard work over the years and I wanted to ask a few of quick questions.

I have been through the entire thread twice and cannot find clear answers, so forgive me if I missed them.

The first relates to a similar question/issue that was raised in the Forward Test thread today. viewtopic.php?f=10&t=284519&start=250#p5413121

So this prompted me to post.

First, at some point you moved from having the bridge LMP for social security from a separate fund to having a "virtual" approach as part of the entire portfolio. This post from the thread shows a good explanation of your thinking. viewtopic.php?f=10&t=120430&start=900#p4879094

This post illustrates the issue I am having in that when you include the LMP bridge in the total portfolio, must you also adjust the asset allocation input (e.g. more bonds) into that parameter on the VPW spreadsheet? When I first put numbers into the spreadsheet they seemed too good to be true. I then worked off just the table, and backed out the bridge LMP, and realized the issue was that I was using a higher stock bond percentage for the entire portfolio. As you are aware, VPW is based on year and asset allocation so the higher stock allocation increased the withdrawal percentage. Looking at examples in other posts you do indeed lower the stock/bond percentage when the bridge LMP is not segregated but held with the total portfolio. But I am not sure this is clear when placing inputs into the VPW spreadsheet.

The second issue is, if you adjust your allocation for the LMP bridge, how does that integrate with the one fund approach you advocate. Assume for example, I purchase a 40/60 Life Strategy Fund at age 62, which includes my bridge. I want my age 70-99 allocation to be in a 60/40 Life Strategy Fund. Am I forced at age 70 to liquidate and reinvest?

The third issue is much simpler--at some point you switched from a 20% terminal withdrawal percentage to 10% and I could not find the rationale for that.

Thanks again for all your work. I have been reading lately on withdrawal methods, and the VPW method should be on anyone's short list. For me, it was quite helpful because I have a fair amount of commercial real estate with variable income, and the VPW allows to me to confidently take more from my portfolio assets earlier in retirement.
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

Lastrun wrote: Tue Aug 04, 2020 5:29 pm First Longinvest I want to thank you for all your hard work over the years and I wanted to ask a few of quick questions.

I have been through the entire thread twice and cannot find clear answers, so forgive me if I missed them.

The first relates to a similar question/issue that was raised in the Forward Test thread today. viewtopic.php?f=10&t=284519&start=250#p5413121

So this prompted me to post.

First, at some point you moved from having the bridge LMP for social security from a separate fund to having a "virtual" approach as part of the entire portfolio. This post from the thread shows a good explanation of your thinking. viewtopic.php?f=10&t=120430&start=900#p4879094

This post illustrates the issue I am having in that when you include the LMP bridge in the total portfolio, must you also adjust the asset allocation input (e.g. more bonds) into that parameter on the VPW spreadsheet? When I first put numbers into the spreadsheet they seemed too good to be true. I then worked off just the table, and backed out the bridge LMP, and realized the issue was that I was using a higher stock bond percentage for the entire portfolio. As you are aware, VPW is based on year and asset allocation so the higher stock allocation increased the withdrawal percentage. Looking at examples in other posts you do indeed lower the stock/bond percentage when the bridge LMP is not segregated but held with the total portfolio. But I am not sure this is clear when placing inputs into the VPW spreadsheet. In other words, I think that using the spreadsheet with a constant allocation is good enough.

The second issue is, if you adjust your allocation for the LMP bridge, how does that integrate with the one fund approach you advocate. Assume for example, I purchase a 40/60 Life Strategy Fund at age 62, which includes my bridge. I want my age 70-99 allocation to be in a 60/40 Life Strategy Fund. Am I forced at age 70 to liquidate and reinvest?

The third issue is much simpler--at some point you switched from a 20% terminal withdrawal percentage to 10% and I could not find the rationale for that.

Thanks again for all your work. I have been reading lately on withdrawal methods, and the VPW method should be on anyone's short list. For me, it was quite helpful because I have a fair amount of commercial real estate with variable income, and the VPW allows to me to confidently take more from my portfolio assets earlier in retirement.
Lastrun, thanks for the nice comments.

Answer to first two questions:

With the new worksheet, things are much simpler. There's a single undivided portfolio with a single asset allocation. There's no bridge portfolio or separate internal allocations. The spreadsheet handles current and delayed pensions, with and without cost-of-living adjustments.

The new worksheet provides an estimate of the impact of a -50% stock loss on total retirement income (including pensions and portfolio withdrawals). That's the important thing to worry about. If the estimated loss or withdrawal reduction is too big, even when delaying Social Security to age 70, there are at least three possible solutions: (1) increase the bond allocation, (2) use a small part of the portfolio to buy an inflation-indexed (e.g. 2%-indexed) SPIA*, or (3) work a few additional years before retiring.

* Single Premium Immediate Annuity.

Answer to third question:

The choice of 10% provides a softer decline in old age (within the VPW theoretical model) of the portfolio balance and withdrawals. It's assumed, of course, that the retiree made sure at age 80 to have sufficient lifelong non-portfolio income to comfortably live independently from portfolio withdrawals. (This is substantially more than only covering bare-bones expenses!) With a 10% maximum percentage, the decline starts around age 90 and reduces the residual portfolio and withdrawals by approximately half (within the model) every 10 years, preserving some liquidity all lifelong.

Here's a comparative chart of theoretical models. In real life, actual portfolio returns are applied on top of these models:

Image
Last edited by longinvest on Fri Aug 28, 2020 10:53 pm, edited 1 time in total.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
Tejfyy
Posts: 106
Joined: Mon Aug 26, 2019 9:18 pm

Re: Variable Percentage Withdrawal (VPW)

Post by Tejfyy »

longinvest wrote: Wed Aug 05, 2020 6:26 am
With the new worksheet, things are much simpler. There's a single undivided portfolio with a single asset allocation. There's no bridge portfolio or separate internal allocations. The spreadsheet handles current and delayed pensions, with and without cost-of-living adjustments.

The new worksheet provides an estimate of the impact of a -50% stock loss on total retirement income (including pensions and portfolio withdrawals). That's the important thing to worry about. If the estimated loss or withdrawal reduction is too big, even when delaying Social Security to age 70, there are at least three possible solutions: (1) increase the bond allocation, (2) use a small part of the portfolio to buy an inflation-indexed (e.g. 2%-indexed) SPIA*, or (3) work a few additional years before retiring.

* Single Premium Immediate Annuity.

Answer to third question:

The choice of 10% provides a softer decline in old age (within the VPW theoretical model) of the portfolio balance and withdrawals. It's assumed, of course, that the retiree made sure at age 80 to have sufficient lifelong non-portfolio income to comfortably live independently from portfolio withdrawals. (This is substantially more than only covering bare-bones expenses!) With a 10% maximum percentage, the decline starts around age 90 and reduces the residual portfolio and withdrawals by approximately half (within the model) every 10 years, preserving some liquidity all lifelong.
I too find it wonderfully simple and I thank you for your dedication and generosity!

How often should you update the retirement worksheet? It says each year, but what if there’s a 50% drop in the middle of the year? Or put another way, are there periods when you should revisit it more frequently, whatever “more frequently” would be? Or, is it most effective when you set it up each year and leave it?

I don't quite understand part of your answer to the 3rd question. With "(This is substantially more than only covering bare-bones expenses!) do you mean your lifetime income i.e. social security and a pension should be substantially more?

Thank you again longinvest!
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

Tejfyy wrote: Sun Aug 09, 2020 9:54 pm How often should you update the retirement worksheet? It says each year, but what if there’s a 50% drop in the middle of the year? Or put another way, are there periods when you should revisit it more frequently, whatever “more frequently” would be? Or, is it most effective when you set it up each year and leave it?
Different retirees have different preferences for the frequency of their portfolio withdrawals. For those who choose to make a single withdrawal per year, I'd suggest updating the worksheet once a year just before withdrawal. For those who choose to make four withdrawals per year, I'd suggest updating the worksheet every quarter just before withdrawal. For those who choose to make twelve withdrawals per year, I'd suggest updating the worksheet every month just before withdrawal.
Tejfyy wrote: Sun Aug 09, 2020 9:54 pm I don't quite understand part of your answer to the 3rd question. With "(This is substantially more than only covering bare-bones expenses!) do you mean your lifetime income i.e. social security and a pension should be substantially more?
I just mean that a "comfortable life" must have "comfort" in it; it's more than barely "surviving". This is in the context of our wiki's explanation of how to use variable percentage withdrawals during retirement:

Steps
3. At age 80, if you're still alive, it's important to consider using part (but not all) of your remaining portfolio to buy an inflation-indexed Single Premium Immediate Annuity (SPIA), so that the estimated Income Floor After 100 is sufficient to live comfortably, independently of future portfolio withdrawals. This aims to reduce the financial risks associated with living past age 100.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
Sandi_k
Posts: 1577
Joined: Sat May 16, 2015 11:55 am
Location: SF Bay Area

Re: Variable Percentage Withdrawal (VPW)

Post by Sandi_k »

@longinvest, I've puttered around with the worksheet, and had a couple of questions.

1) I already have a pension, but your "Buy an SPIA at age 80" comes up regardless. You might want to say "If you do not have a COLA-adjusted pension aside from Social Security, you should consider purchasing an SPIA at age 80." ;)

2) On the accumulation tab: I have Soc Security listed as Pension #1 (me) and Pension #2 (DH). Should we be adding in our OASDI contributions to the spreadsheet in the grayed out cell for "Pension #X contribution?"

Thanks!
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

Sandi,
Sandi_k wrote: Mon Aug 10, 2020 3:12 pm 1) I already have a pension, but your "Buy an SPIA at age 80" comes up regardless. You might want to say "If you do not have a COLA-adjusted pension aside from Social Security, you should consider purchasing an SPIA at age 80." ;)
The Instructions sheet of the VPW Worksheet says:
Retirement Sheet
...
- Starting at age 80, the sheet displays the projected income floor after age 100.
...
Starting at age 80, the Retirement sheet of the VPW Worksheet displays an additional "Income Floor After Age 100 *" cell. The asterisk refers to a footnote which also starts displaying at age 80:
* Can be increased by buying an inflation-indexed SPIA** with part of the portfolio.
** Single-Premium Immediate Annuity.
As for our Wiki, when explaining how to use the VPW Worksheet, it says:
Steps
3. At age 80, if you're still alive, it's important to consider using part (but not all) of your remaining portfolio to buy an inflation-indexed Single Premium Immediate Annuity (SPIA), so that the estimated Income Floor After 100 is sufficient to live comfortably, independently of future portfolio withdrawals. This aims to reduce the financial risks associated with living past age 100.
The key word, here, is "consider". When the income floor after age 100 is already sufficient to live comfortably (independently of future portfolio withdrawals), it seems clear that there's no need to buy anything.

Sandi_k wrote: Mon Aug 10, 2020 3:12 pm 2) On the accumulation tab: I have Soc Security listed as Pension #1 (me) and Pension #2 (DH). Should we be adding in our OASDI contributions to the spreadsheet in the grayed out cell for "Pension #X contribution?"
Effectively, OASDI contributions should be included as a pension contribution in the Accumulation sheet. (See this thread for detailed explanations of the Accumulation sheet).

When planning retirement for a couple, there's a tricky issue with pensions that aren't 100% joint and survivor. It's important to consider what happens when one spouse dies. In particular, the smallest of the two Social Security payment streams stops (or never starts, if the spouse dies early). Depending on the financial situation of the couple, this might represent a significant drop in retirement income for the survivor.

One planning approach is to simply accept that the survivor will have to live on a smaller survivor income. When necessary, the survivor could move to a less expensive home, etc.

Another planning approach is to only include the survivor pension amount, in the worksheet, instead of the full amount when both are alive. This eliminates the income drop on the first death, but it's likely to lead to extra (unplanned for) income when pensions start and for as long as both are alive. This extra income could be used for gifts, unplanned discretionary expenses, etc.

A middle-way planning approach is to aim for a smaller survivor income drop by including a higher amount than the survivor pension, but a smaller amount than the full pension amount (when both are alive) in the worksheet, reducing the size of the income drop on the first death.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
Sandi_k
Posts: 1577
Joined: Sat May 16, 2015 11:55 am
Location: SF Bay Area

Re: Variable Percentage Withdrawal (VPW)

Post by Sandi_k »

Thank you!

I am electing for 100% survivor benefits for DH; it reduces the monthly benefit by ~ 14%, but seems cheap given it's so generous, and the COLA matters over 3 decades.

Social Security for both of us I have modeled in a "flat" spreadsheet; assuming DH goes first, there will be more than enough in the remaining SocSec and pension income for a singleton. The tax hit will be sad, but we'd rather model higher income and higher spend while we're both here to enjoy it. ;)
Stubbie
Posts: 27
Joined: Fri May 29, 2020 6:09 am

Re: Variable Percentage Withdrawal (VPW)

Post by Stubbie »

longinvest wrote: Mon Aug 10, 2020 5:23 pm Another planning approach is to only include the survivor pension amount, in the worksheet, instead of the full amount when both are alive. This eliminates the income drop on the first death, but it's likely to lead to extra (unplanned for) income when pensions start and for as long as both are alive. This extra income could be used for gifts, unplanned discretionary expenses, etc.
This is the approach that we have chosen, knowing that we likely could spend more money each year than the spreadsheet indicates, but also knowing that there is added ability to deal with future unplanned financial calamities.

Longinvest, let me add my thanks to you for spending the time to create such a great tool!

Stub
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

In the VPW forward test thread, forum members Azanon and Hitchcock_Fan made an improvement suggestion for the VPW Accumulation and Retirement Worksheet:
azanon wrote: Mon Aug 24, 2020 10:22 am Finally got a chance to play with the "new" VWP- Accumulationo-and-retirement-worksheet. I have a wishlist item for the accumulation and retirement tab: I wish there were an end date (by age) option for the various defined benefit pension entries. Not all "pensions" (or pension variants) are for life, and I'll have one that has an ending date (ends at age 62, specifically).

Love it otherwise though - thanks so much!
Hitchcock_Fan wrote: Mon Aug 24, 2020 8:14 pm I'll second this emotion. I'm in same situation with a pension variant that ends at age 62. I usually just leave it out of calculations, but it would be nice to include it.
There was also a suggestion on our Canadian sister forum to handle part-time, casual, or business income for a few years.

As a result of these interesting suggestions, I'm considering to add "Temporary Retirement Income" blocks at the bottom of both Accumulation and Retirement worksheets, relatively similar to "Defined Benefit Pension" blocks. These blocks could be used for part-time work income during retirement and for temporary pensions.

What do you think?
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
retiringwhen
Posts: 1979
Joined: Sat Jul 08, 2017 10:09 am
Location: New Jersey, USA

Re: Variable Percentage Withdrawal (VPW)

Post by retiringwhen »

longinvest wrote: Tue Aug 25, 2020 7:36 am As a result of these interesting suggestions, I'm considering to add "Temporary Retirement Income" blocks at the bottom of both Accumulation and Retirement worksheets, relatively similar to "Defined Benefit Pension" blocks. These blocks could be used for part-time work income during retirement and for temporary pensions.

What do you think?
I like the idea, it can be used for two other purposes folks may wish to model.

1.) lower earning spouse's SS benefits from initial benefit until they convert to full spousal benefits.
2.) model potential outcomes where one spouse dies and their portion of SS ends.

These are more for what-if scenarios but can be useful. I have been playing with Mike Piper's https://opensocialsecurity.com/ and taking the outcomes there and modelling them with VPW can help understand how they interact.
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

LakerP wrote: Tue Aug 25, 2020 2:26 am I'm simply having a hard time believing this spreadsheet whatsoever, putting in my numbers as a 25 year old with a million dollars invested, it tells me I should start withdrawing $43,717 for 2020 in a 80/20 Stocks/Bonds portfolio. This is obviously an utterly insane suggestion to make, if this spreadsheet doesn't work for all age groups/circumstances then I have a hard time justifying its use for anyone else...
LakerP, the variable percentage withdrawal (VPW) method works differently from constant-dollar withdrawal also known as "safe" withdrawal rate (SWR).

The VPW Retirement Worksheet suggests a variable withdrawal amount and it's explicit about it. While it suggests a withdrawal amount for the current year, it also informs about the required flexibility to handle the consequences of stocks losing -50% of their value. This isn't a maximum loss; it's a normal loss for stocks that shouldn't affect the retiree's comfort in any way.

I've been able to reproduce the $43,717 suggestion for the current year using the following parameters: age 25, $1,000,000 portfolio, 80/20 stock/bond allocation, and no Social Security. The VPW Retirement Worksheet explicitly indicates that the retiree must have the flexibility to reduce withdrawals by -$17,487, down to a $26,230 annual income while suffering a -$400,000 portfolio loss. Here's a screenshot:

Image

VPW is about planning to live comfortably with a variable income during retirement even when stocks lose -50% of their value.

It's my opinion that there's nothing comfortable about ending up with only $2,185/month for taxes and expenses with no employment and no employer health-care plan after a severe bear market.

My suggestion for a 25 years old person is to consider using the VPW Accumulation Worksheet, instead. It's very helpful to determine reasonable (variable) retirement portfolio contribution amounts, during accumulation, to balance quality of life during both accumulation and retirement. Detailed explanations are provided in this thread.
Last edited by longinvest on Fri Aug 28, 2020 10:54 pm, edited 1 time in total.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
Lastrun
Posts: 345
Joined: Wed May 03, 2017 6:46 pm

Re: Variable Percentage Withdrawal (VPW)

Post by Lastrun »

longinvest wrote: Tue Aug 25, 2020 7:36 am In the VPW forward test thread, forum members Azanon and Hitchcock_Fan made an improvement suggestion for the VPW Accumulation and Retirement Worksheet:

What do you think?
I agree, to me it would be helpful with those that have term certain pensions, SPIAs or an LMP designed only to provide for a term of years.
MrDrinkingWater
Posts: 193
Joined: Tue Dec 23, 2014 11:30 am

Re: Variable Percentage Withdrawal (VPW)

Post by MrDrinkingWater »

longinvest wrote: Tue Aug 25, 2020 7:36 am In the VPW forward test thread, forum members Azanon and Hitchcock_Fan made an improvement suggestion for the VPW Accumulation and Retirement Worksheet:
azanon wrote: Mon Aug 24, 2020 10:22 am Finally got a chance to play with the "new" VWP- Accumulationo-and-retirement-worksheet. I have a wishlist item for the accumulation and retirement tab: I wish there were an end date (by age) option for the various defined benefit pension entries. Not all "pensions" (or pension variants) are for life, and I'll have one that has an ending date (ends at age 62, specifically).

Love it otherwise though - thanks so much!
Hitchcock_Fan wrote: Mon Aug 24, 2020 8:14 pm I'll second this emotion. I'm in same situation with a pension variant that ends at age 62. I usually just leave it out of calculations, but it would be nice to include it.
There was also a suggestion on our Canadian sister forum to handle part-time, casual, or business income for a few years.

As a result of these interesting suggestions, I'm considering to add "Temporary Retirement Income" blocks at the bottom of both Accumulation and Retirement worksheets, relatively similar to "Defined Benefit Pension" blocks. These blocks could be used for part-time work income during retirement and for temporary pensions.

What do you think?
I like the idea of adding in this proposed new "Temporary Retirement Income" feature.

Those who have limited-duration pensions and legal settlements that flow into their retirement assets and spending plans would definitely benefit. Some folks also will be taking on intermittent, limited-duration, project-oriented work from time to time for a few weeks per year. I think many of those VPW users were probably just "ball-parking" in those additional inflows before.

Would there be any advantage to making it a "+" or "-" feature (limited-term income or limited-term expense)? Would knowing that something was a limited term expense help estimate a better Annual Portfolio Withdrawal Reduction suggestion for the current year?
SevenBridgesRoad
Posts: 739
Joined: Sat Jul 07, 2018 12:14 am

Re: Variable Percentage Withdrawal (VPW)

Post by SevenBridgesRoad »

retiringwhen wrote: Tue Aug 25, 2020 8:45 am
longinvest wrote: Tue Aug 25, 2020 7:36 am As a result of these interesting suggestions, I'm considering to add "Temporary Retirement Income" blocks at the bottom of both Accumulation and Retirement worksheets, relatively similar to "Defined Benefit Pension" blocks. These blocks could be used for part-time work income during retirement and for temporary pensions.

What do you think?
I like the idea...

...lower earning spouse's SS benefits from initial benefit until they convert to full spousal benefits...
Yes, as a retired user of the worksheet, I like the idea of this proposed feature. I used a workaround to show my spouse drawing her benefit now, while I delay mine until age 70 and she converts to the higher spousal benefit at that point. But the feature you are considering would make it more clear. Thanks!
KarenC
Posts: 163
Joined: Mon Apr 27, 2015 7:25 am

Re: Variable Percentage Withdrawal (VPW)

Post by KarenC »

MrDrinkingWater wrote: Tue Aug 25, 2020 12:29 pm […]

Would there be any advantage to making it a "+" or "-" feature (limited-term income or limited-term expense)? Would knowing that something was a limited term expense help estimate a better Annual Portfolio Withdrawal Reduction suggestion for the current year?
I would like that. I had to roll my own version of VPW to model increased healthcare expenses before Medicare.
"The first principle is that you must not fool yourself, and you are the easiest person to fool." — Richard P. Feynman
User avatar
4nursebee
Posts: 1653
Joined: Sun Apr 01, 2012 7:56 am
Location: US

Re: Variable Percentage Withdrawal (VPW)

Post by 4nursebee »

I finally figured a bit out on the VPW tables in that they are two different spreadsheets to download, glad that got straight in my head. I've looked this over many times trying to backtest. Can the backtest not be made part of or linked with the other spreadsheets?

The accumulation retirement spreadsheet does not allow an input of 100% stocks whereas the backtest one does. Can the retirement sheet be updated to include 100$ stocks?

Easily found tables of suggested withdrawal rates using VPW do not go all the way up to 100% stocks, does such a table exist? This is the table I see easily: https://www.bogleheads.org/wiki/Variabl ... withdrawal

For running backtest scenarios, what are the worst years to start with? Best years?

Thank you for everyone that has worked on this tool and method.
4RNB
Pale Blue Dot
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

4nursebee wrote: Thu Aug 27, 2020 5:00 am I finally figured a bit out on the VPW tables in that they are two different spreadsheets to download, glad that got straight in my head. I've looked this over many times trying to backtest. Can the backtest not be made part of or linked with the other spreadsheets?

The accumulation retirement spreadsheet does not allow an input of 100% stocks whereas the backtest one does. Can the retirement sheet be updated to include 100$ stocks?

Easily found tables of suggested withdrawal rates using VPW do not go all the way up to 100% stocks, does such a table exist? This is the table I see easily: https://www.bogleheads.org/wiki/Variabl ... withdrawal

For running backtest scenarios, what are the worst years to start with? Best years?

Thank you for everyone that has worked on this tool and method.
4RNB
4nursebee, the VPW backtesting spreadsheet was used to develop VPW and can be used to understand how it would have worked in the past. In contrast, the goal of the new VPW accumulation and retirement worksheet is to be an easy-to-use tool during accumulation (to determine portfolio contribution amounts) and retirement (to determine portfolio withdrawal amounts) while taking into account pensions (including Social Security) and dampening longevity-related financial risks.

As your goal is to understand how VPW would have reacted to various market returns, you should use the backtesting spreadsheet. Its "Withdrawal Statistics for Every Start Year" chart provides the information you're seeking (e.g. worst/best start years).

One must be careful not to project the past into the future. Backtesting is an excellent tool to discover previously unknown risks, but it can only reveal the few risks that showed up in the little available historical data available to us. In real life, it's important to build a robust financial plan which doesn't depend on the future being no worse than the past. Our wiki's VPW page strongly suggests to combine stable lifelong non-portfolio income with variable withdrawals from a balanced (stocks and bonds) portfolio which adapt to actual returns during retirement.
Last edited by longinvest on Thu Aug 27, 2020 7:02 am, edited 1 time in total.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
4nursebee
Posts: 1653
Joined: Sun Apr 01, 2012 7:56 am
Location: US

Re: Variable Percentage Withdrawal (VPW)

Post by 4nursebee »

longinvest wrote: Thu Aug 27, 2020 6:34 am
4nursebee wrote: Thu Aug 27, 2020 5:00 am I finally figured a bit out on the VPW tables in that they are two different spreadsheets to download, glad that got straight in my head. I've looked this over many times trying to backtest. Can the backtest not be made part of or linked with the other spreadsheets?

The accumulation retirement spreadsheet does not allow an input of 100% stocks whereas the backtest one does. Can the retirement sheet be updated to include 100$ stocks?

Easily found tables of suggested withdrawal rates using VPW do not go all the way up to 100% stocks, does such a table exist? This is the table I see easily: https://www.bogleheads.org/wiki/Variabl ... withdrawal

For running backtest scenarios, what are the worst years to start with? Best years?

Thank you for everyone that has worked on this tool and method.
4RNB
4nursebee, the VPW backtesting spreadsheet was used to develop VPW and can be used to understand how it would have worked in the past. In contrast, the goal of the new VPW accumulation and retirement worksheet is to be an easy-to-use tool during accumulation (to determine portfolio contribution amounts) and retirement (to determine portfolio withdrawal amounts) while taking into account pensions (including Social Security) and dampening longevity-related financial risks.

As your goal is to understand how VPW would have reacted to various market returns, you should use the backtesting spreadsheet. Its "Withdrawal Statistics for Every Start Year" chart provides the information you're seeking (e.g. worst/best start years).
OK, I think I found that chart, right side, 2nd item down. It does not seem to adjust for our inputs, is it for a default setting of say 1M? Same thing with the back testing charts on the top, no change with inputs?
Pale Blue Dot
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

4nursebee wrote: Thu Aug 27, 2020 7:00 am
longinvest wrote: Thu Aug 27, 2020 6:34 am As your goal is to understand how VPW would have reacted to various market returns, you should use the backtesting spreadsheet. Its "Withdrawal Statistics for Every Start Year" chart provides the information you're seeking (e.g. worst/best start years).
OK, I think I found that chart, right side, 2nd item down. It does not seem to adjust for our inputs, is it for a default setting of say 1M? Same thing with the back testing charts on the top, no change with inputs?
The chart automatically adapts to the various parameter (start and last withdrawal age, initial portfolio size, asset allocation, data set).
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
4nursebee
Posts: 1653
Joined: Sun Apr 01, 2012 7:56 am
Location: US

Re: Variable Percentage Withdrawal (VPW)

Post by 4nursebee »

longinvest wrote: Thu Aug 27, 2020 7:09 am
4nursebee wrote: Thu Aug 27, 2020 7:00 am
longinvest wrote: Thu Aug 27, 2020 6:34 am As your goal is to understand how VPW would have reacted to various market returns, you should use the backtesting spreadsheet. Its "Withdrawal Statistics for Every Start Year" chart provides the information you're seeking (e.g. worst/best start years).
OK, I think I found that chart, right side, 2nd item down. It does not seem to adjust for our inputs, is it for a default setting of say 1M? Same thing with the back testing charts on the top, no change with inputs?
The chart automatically adapts to the various parameter (start and last withdrawal age, initial portfolio size, asset allocation, data set).
For me the tables adjust but not the charts. I tried to PM you to share screen capture or the google spreadsheet but you seem to have that off.

I changed the portfolio start to 300K, nothing changed, image should be to show what I see.
https://i.postimg.cc/hGHbpXhy/LI2.png
Pale Blue Dot
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

4nursebee wrote: Thu Aug 27, 2020 7:12 am
longinvest wrote: Thu Aug 27, 2020 7:09 am The chart automatically adapts to the various parameter (start and last withdrawal age, initial portfolio size, asset allocation, data set).
For me the tables adjust but not the charts. I tried to PM you to share screen capture or the google spreadsheet but you seem to have that off.

I changed the portfolio start to 300K, nothing changed, image should be to show what I see.
https://i.postimg.cc/hGHbpXhy/LI2.png
You seem to be using the online Google Sheets version. It is very slow at making the numerous calculations for all start years and updating that chart. In other words, it requires patience. The original LibreOffice Calc version (*.ods) and the Microsoft Excel version (*.xlsx) are much faster on a computer.
Last edited by longinvest on Thu Aug 27, 2020 7:59 am, edited 1 time in total.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
4nursebee
Posts: 1653
Joined: Sun Apr 01, 2012 7:56 am
Location: US

Re: Variable Percentage Withdrawal (VPW)

Post by 4nursebee »

longinvest wrote: Thu Aug 27, 2020 7:31 am
4nursebee wrote: Thu Aug 27, 2020 7:12 am
longinvest wrote: Thu Aug 27, 2020 7:09 am The chart automatically adapts to the various parameter (start and last withdrawal age, initial portfolio size, asset allocation, data set).
For me the tables adjust but not the charts. I tried to PM you to share screen capture or the google spreadsheet but you seem to have that off.

I changed the portfolio start to 300K, nothing changed, image should be to show what I see.
https://i.postimg.cc/hGHbpXhy/LI2.png
You seem to be using the online Google Sheet version. It is very slow at making the numerous calculations for all start years and updating that chart. In other words, it requires patience. The original LibreOffice Calc version (*.ods) and the Microsoft Excel version (*.xlsx) are much faster on a computer.
OK, but I have had the window open since last posting and the calculations have not populated yet. Not needed though.
Thanks again for your work
Pale Blue Dot
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

4nursebee wrote: Thu Aug 27, 2020 7:36 am OK, but I have had the window open since last posting and the calculations have not populated yet.
Actually, I've just noticed that somehow all charts in the Google Sheets version had been replaced with static images! I've fixed that and put back dynamic charts. Could you make a new copy and try it again?
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
4nursebee
Posts: 1653
Joined: Sun Apr 01, 2012 7:56 am
Location: US

Re: Variable Percentage Withdrawal (VPW)

Post by 4nursebee »

longinvest wrote: Thu Aug 27, 2020 7:51 am
4nursebee wrote: Thu Aug 27, 2020 7:36 am OK, but I have had the window open since last posting and the calculations have not populated yet.
Actually, I've just noticed that somehow all charts in the Google Sheets version had been replaced with static images! I've fixed that and put back dynamic charts. Could you make a new copy and try it again?
:sharebeer
Looks good, thanks
Pale Blue Dot
Seasonal
Posts: 2269
Joined: Sun May 21, 2017 1:49 pm

Re: Variable Percentage Withdrawal (VPW)

Post by Seasonal »

Sorry for asking things that have no doubt been covered earlier, but, nonetheless, is the following correct?

It appears the key calculation is a the annual withdrawal rate, which is calculated with Excel: pmt(rate, years, -1, 0, 1), where rate is based on 5% for stocks and 1.8% for bonds and years is the number of years to deplete the portfolio. The result is then multiplied by remaining balance to get annual spending. There's a lot more in the VPW and Table tabs, but they could be essentially replaced with this calculation.

The spreadsheet asks for domestic/international stock allocation, but only uses the total percent.
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

Seasonal wrote: Thu Aug 27, 2020 9:55 am It appears the key calculation is a the annual withdrawal rate, which is calculated with Excel: pmt(rate, years, -1, 0, 1), [...].
Seasonal, the mathematics behind the VPW Table are effectively quite simple. But, there's more to VPW than these simple mathematics; there's an entire accumulation and retirement approach using variable portfolio contributions and withdrawals (without attempting to predict future returns) while taking pensions into account and addressing longevity concerns.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
Seasonal
Posts: 2269
Joined: Sun May 21, 2017 1:49 pm

Re: Variable Percentage Withdrawal (VPW)

Post by Seasonal »

longinvest wrote: Thu Aug 27, 2020 5:56 pm
Seasonal wrote: Thu Aug 27, 2020 9:55 am It appears the key calculation is a the annual withdrawal rate, which is calculated with Excel: pmt(rate, years, -1, 0, 1), [...].
Seasonal, the mathematics behind the VPW Table are effectively quite simple. But, there's more to VPW than these simple mathematics; there's an entire accumulation and retirement approach using variable portfolio contributions and withdrawals (without attempting to predict future returns) while taking pensions into account and addressing longevity concerns.
Yes, and your Forward Test thread is an excellent illustration. I just wanted to make sure I understood the math function, which is essentially to run PMT to compute the annual portfolio withdrawal, which in turn is based on a default of 5% for stocks and 1.8% for bonds.
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

Seasonal wrote: Fri Aug 28, 2020 4:09 am I just wanted to make sure I understood the math function, which is essentially to run PMT to compute the annual portfolio withdrawal, which in turn is based on a default of 5% for stocks and 1.8% for bonds.
Seasonal, the stock and bond growth trends are respectively 5% and 1.9%. They're based on the long-term returns of world stocks and bonds from 1900 to 2018 according the Summary Edition of the Credit Suisse Global Investment Returns Yearbook 2019.

A growth trend is a timeless wild-ass guess that aims to represent an annual return that is lower than a high annual return and higher than a low annual return. Its role is to distinguish between a high annual return and a low annual return. It's not a prediction of future returns. It is fixed and must never be changed. At the top of a bubble, it is likely to be higher than future returns. At the bottom of a crash it is likely to be lower than future returns.

The use of fixed growth trends, instead of variable future return predictions, is fundamental to VPW. Here's part of a previous post I wrote about it:
longinvest wrote: Sun Jun 16, 2019 4:00 pm The VPW method is a Bogleheads method that never tries to time the market.

Many pundits like to make 10-year return predictions to gain media exposure. The financial media loves such predictions. Yet, it's safe enough for pundits; almost nobody will remember their predictions in 10 years. Do you really remember a prediction made by a pundit 10 years ago?

Let's do a mental exercise. Let's assume, for a second, that we had a perfect return prediction for the next 10 years. Remember that the growth trend distinguishes between high and low returns. The thing is this: even if the future return of next 10 years is low, it still doesn't make sense to use it as growth trend. The growth trend isn't a future return prediction.

It's important to think ahead of time about the consequences of one's choices. Decreasing withdrawals (because future returns are low) when the portfolio is up and increasing withdrawals (because future returns are high) when the portfolio is down is a "sell less high, sell more low" approach to portfolio withdrawals. It's a form of "buy high, sell low", a counterproductive investing approach. It wouldn't be a good idea to do that.

The VPW method sensibly ignores all future return predictions.
Last edited by longinvest on Fri Aug 28, 2020 6:34 am, edited 1 time in total.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
Seasonal
Posts: 2269
Joined: Sun May 21, 2017 1:49 pm

Re: Variable Percentage Withdrawal (VPW)

Post by Seasonal »

Thanks - I was looking at an older version of the spreadsheet which used the Credit Suisse Global Investment Returns Yearbook 2016. It was VPW version 2.1.

I really should keep current.

EDIT:

The new version is very different from the version I was looking at. Apologies for the confusion.
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

Seasonal wrote: Fri Aug 28, 2020 6:34 am Thanks - I was looking at an older version of the spreadsheet which used the Credit Suisse Global Investment Returns Yearbook 2016.

I really should keep current.
Seasonal, I wouldn't worry about it; there's no need to update anything. What you've got is good enough. If you want the exact numbers, the trends in that older version were ((300 ^ (1 / 116)) - 1) [ = 5.04% ] for stocks and ((8 ^ (1 / 116)) - 1) [ = 1.81% ] for bonds. The difference is small and of little consequence; growth trends are just wild ass guesses. There was a previous discussion about whether growth trends should be updated once in a while in the spreadsheet:
longinvest wrote: Fri Jul 20, 2018 8:43 pm Whether the specific percentage in the VPW table is 4.8% or 4.9% isn't important. These are rounded values resulting from a calculation based on a wild-ass guess. What's important is to understand that it wouldn't be sustainable to withdraw 7% or 8% instead of 4.8%, and keep over-withdrawing like that regularly. I've kept one decimal of precision (after being rightfully chastised by forum member Rodc, early in this thread, for the false precision of using the default two decimals of Microsoft Excel/Libreoffice Calc) so that annual percentage adjustments remain small. The difference between 4% and 5% of a portfolio is big; the difference between 4.4% and 4.5% is much smaller and leads to a more reasonable annual adjustment.
[...]
I have no strong opinion about whether to update the spreadsheet's trends based on the latest Credit Suisse Yearbook Summary or not, except that I want users to know that they can take a single copy of the VPW table and use it all retirement long.

A compromise would be to update the trends every 5 years (in 2020, 2025, etc) just to reinforce the idea of stability; that there's no need to update the VPW table yearly (or ever). Would that be more intellectually satisfying than using stale values from 2016?
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

In another thread, forum member Midareff wrote:
midareff wrote: Sun Aug 09, 2020 6:41 am I like the VPW as a guideline with a small change in that I use the application of the average of the three prior years portfolio balance for smoothing year to year undulations.
Unfortunately, smoothing portfolio withdrawals (e.g. taking a withdrawal from the portfolio which is bigger than what the plain VPW calculation indicates when the portfolio is down in value) hurts the portfolio, especially during a severe bear market. It's something to avoid.

It can be useful to distinguish the concept of portfolio withdrawal from the concept of income from portfolio withdrawals. It's easy to deliver smoothed income without smoothing portfolio withdrawals by simply putting a few months of withdrawals into a high-interest savings account (called withdrawal cushion).

The VPW forward test thread uses a 5-month withdrawal cushion. Here's how it works:
  • The retiree initially puts 5 months of withdrawals into the withdrawal cushion (a high-interest savings account).
  • Each month:
    1. The retiree withdraws 1/12 of (Portfolio balance X VPW Table percentage).
    2. The retiree adds the withdrawn amount into the withdrawal cushion, increasing it to 6 months of withdrawals.
    3. The retiree immediately transfers 1/6 of the withdrawal cushion to the monthly spending bank account (reducing it back to 5 months of withdrawals).
Steps 2 and 3 result into smoothed monthly income, even though monthly withdrawals aren't smoothed. The two steps are usually combined into a single money transfer to or from the withdrawal cushion using arithmetic calculations.

Earlier in this thread, I've calculated that:
longinvest wrote: Sat Jul 04, 2020 7:56 pm The composition of [annual income from 12 monthly withdrawals when using a 5-month withdrawal cushion] is this [...]: 63% current year, (37% X 90%) = 33% previous year, and 4% 2 years ago.
* modified for clarity.

Using the formulas explained in the linked post, I get that the composition of annual income from 12 monthly withdrawals when using a 15-month withdrawal cushion is: 33% current year, 36% previous year, and 31% earlier years. That's not perfectly 33%, 33%, and 33%*, but it's close enough and it can be achieved without smoothing withdrawals, just by using a savings account containing 15 months of withdrawals.

It would work like this:
  • The retiree initially puts 15 months of withdrawals into the withdrawal cushion (a high-interest savings account).
  • Each month:
    1. The retiree withdraws 1/12 of (Portfolio balance X VPW Table percentage).
    2. The retiree adds the withdrawn amount into the withdrawal cushion, increasing it to approximately 16 months of withdrawals.
    3. The retiree transfers 1/16 of the withdrawal cushion to the monthly spending bank account (reducing it back to 15 months of withdrawals).
Smoothing income from portfolio withdrawals using a withdrawal cushion is easy and effective. There's no need to hurt the portfolio in bad times with smoothed withdrawals.

Added: Here are links to two posts which provide additional information:
  • The withdrawal cushion dampens short-term fluctuations, bonds dampen deeper and longer fluctuations: this post.
  • Unlike a 5-month withdrawal cushion, bigger withdrawal cushions (such as 15 months) introduce a cash drag (the visual analysis is in the second section of the linked post): this post.
Last edited by longinvest on Sat Aug 29, 2020 3:09 pm, edited 4 times in total.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
dogagility
Posts: 1292
Joined: Fri Feb 24, 2017 6:41 am

Re: Variable Percentage Withdrawal (VPW)

Post by dogagility »

longinvest wrote: Fri Aug 28, 2020 6:21 pm Smoothing income from portfolio withdrawals using a withdrawal cushion is easy and effective. There's no need to hurt the portfolio in bad times with smoothed withdrawals.
Clever! Thanks for all your effort, longinvest. My reasons for using VPW when I "retire" keep growing. :beer
All children spill milk. Learn to smile and wipe it up. -- A Farmer's Wife
User avatar
midareff
Posts: 7324
Joined: Mon Nov 29, 2010 10:43 am
Location: Biscayne Bay, South Florida

Re: Variable Percentage Withdrawal (VPW)

Post by midareff »

longinvest wrote: Fri Aug 28, 2020 6:21 pm In another thread, forum member Midareff wrote:
midareff wrote: Sun Aug 09, 2020 6:41 am I like the VPW as a guideline with a small change in that I use the application of the average of the three prior years portfolio balance for smoothing year to year undulations.
Unfortunately, smoothing portfolio withdrawals (e.g. taking a withdrawal from the portfolio which is bigger than what the plain VPW calculation indicates when the portfolio is down in value) hurts the portfolio, especially during a severe bear market. It's something to avoid.

Unfortunately, without knowing my established income floor and safeguards, it may be easy to assume I would do something silly like aggressive withdrawals during a severe bull market instead of simply smoothing withdrawals for minor variations. We are all getting very spoiled by this bull... do not let recency bias interfere with proper judgement.
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

I should add that a withdrawal cushion only reduces the impact of short-term portfolio fluctuations on income from portfolio withdrawals. It doesn't replace the role of bonds to dampen deeper and longer stock fluctuations. This was explained this post.

Sometimes an image helps (it's extracted from the linked post). The red line represents monthly withdrawals. The green line represents monthly income using a withdrawal cushion. We see that short-term fluctuations are significantly reduced. But, the deeper fluctuations (the green line going up and down over longer time frames) are determined by portfolio allocation.

Image
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
4nursebee
Posts: 1653
Joined: Sun Apr 01, 2012 7:56 am
Location: US

Re: Variable Percentage Withdrawal (VPW)

Post by 4nursebee »

longinvest wrote: Fri Aug 28, 2020 7:02 pm I should add that a withdrawal cushion only reduces the impact of short-term portfolio fluctuations on income from portfolio withdrawals. It doesn't replace the role of bonds to dampen deeper and longer stock fluctuations. This was explained this post.

Sometimes an image helps (it's extracted from the linked post). The red line represents monthly withdrawals. The green line represents monthly income using a withdrawal cushion. We see that short-term fluctuations are significantly reduced. But, the deeper fluctuations (the green line going up and down over longer time frames) are determined by portfolio allocation.

Image
I am being a bit lazy here in that rather read through all the other threads linked I just quickly perused them.
Longinvest, what is the withdrawal cushion and how do you advocate setting up withdrawals using VPW?

More specifically, a few years ago when we stumbled into VPW it inherently made sense, especially if we could build up a reasonable cash reserve. Previously we kept a couple months expenses in the bank, decided pre retirement to build this to 2 years of funds. I think of this two years of funds giving us a cushion in low withdrawal years allowing us to ride out bad market times. We would plan to follow the formula and take out larger amounts in good times, perhaps not use it all.

Our situation:
Fixed income: Pension #1 now, an annuity until first of us reaches age 70, and real estate cover a lot of our expenses. In the future real estate will get paid off and that income will double, SS will kick in for both of us, 2nd pension.
Assets: Cash (non retirement acct, still invested, aka emergency fund) has grown to >7x. Roths might be 28 x. IRA/401K at least 30X. There is another annuity we need to set up to pay out at some time in the future, currently 8x, advised to set this up to pay out last.
Spouse in 1st year retirement, I've just gone part time. We did a Roth conversion, not yet to top of tax bracket, perhaps another 50K to go. My plan would be to follow VPW fully starting next year. I would add up all assets, calculate a withdrawal amount, withdraw that amount ( or a monthly equivalent?) in January. And to be clear, I would calculate withdrawal amount and subtract out the annuity payment first. Otherwise I could base withdrawal amount on non annuity funds. Withdraw for me is a bit of just mental accounting as I'd pull money from the IRA/401K portion, add it to cash portion, pay bills from that. I plan to pull from Roths last. By pulling heavy from tax deferred accounts first I'd hope to take advantage of historically low tax rates, decrease RMDs in the future, help make SS less taxed. I could just pull what we need which would be much less but think this would impact us long term. I might even convert more to a Roth this year or else pull a bit more out and add it to "cash" bucket this year to take us to top of tax bracket. I think some would argue to delay withdrawals, spend from cash bucket first, but we have been advised that our taxes will likely grow in retirement and seek to minimize this.

Questions:
What withdrawal method do you favor, the monthly method?
Does my plan sound reasonable above?
With a mix of assets, what amounts should be used to calculate starting portfolio and VPW? I assume either not using the annuity payout of value is good. What about cash or emergency funds and Roths?
Am I missing something? Do I need to clarify anything?

Thank you.
Pale Blue Dot
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

4nursebee wrote: Sat Aug 29, 2020 2:46 am I am being a bit lazy here in that rather read through all the other threads linked I just quickly perused them.
Longinvest, what is the withdrawal cushion and how do you advocate setting up withdrawals using VPW?
4nursebee,

Monthly withdrawals and the withdrawal cushion are explained in the following posts:
  • Monthly withdrawals and smoothed income (monthly silos): explains the advantages of monthly withdrawals and how to smooth income without introducing cash drag relative to annual withdrawals.
  • A simpler income smoothing approach: while the idea of smoothing income (instead of smoothing withdrawals) was quickly accepted, the use of 12 monthly silos attracted a lot of criticism. A simpler alternative (later called "withdrawal cushion") based on a proposal of forum member D-Dog (and others) was adopted.
  • The decision to consider the withdrawal cushion as part of the total portfolio, for VPW withdrawal calculation purpose, was discussed and subjected to a community vote. Here's the post explaining the issue, my personal opinion, and the result of the vote.
  • A mathematical analysis of the composition of income from withdrawal cushion was provided in this post and this post.
  • An ongoing detailed illustration of monthly withdrawals with a 5-month withdrawal cushion is provided in the VPW forward test thread.
4nursebee wrote: Sat Aug 29, 2020 2:46 am More specifically, a few years ago when we stumbled into VPW it inherently made sense, especially if we could build up a reasonable cash reserve. Previously we kept a couple months expenses in the bank, decided pre retirement to build this to 2 years of funds. I think of this two years of funds giving us a cushion in low withdrawal years allowing us to ride out bad market times. We would plan to follow the formula and take out larger amounts in good times, perhaps not use it all.
(I added the emphasis).

I recently suggested a 15-months withdrawal cushion as an alternative to a member who intended to smooth withdrawals with a 3-year average. Unfortunately, I forgot to discuss its implied cash drag.

The default suggestion is for a 5-month withdrawal cushion because it doesn't introduce a cash drag relative to annual withdrawals. Larger withdrawal cushions do, unfortunately, introduce a cash drag.

Here's a quick visual analysis of cash drag on a model (e.g. before applying real-life market returns) of monthly income from VPW withdrawals. It starts at age 65 with a $1,000,000 portfolio, caps the annual VPW percentage to 10% in old age, and assumes a 60/40 stock/bond allocation. I've included a 23-month withdrawal cushion as it has a similar size to your 2-year cash buffer.
  • Annual withdrawals: A single withdrawal is taken at the beginning of the year. The money is put into a savings account (exactly matching the rate of inflation) and 1/12 is taken as income in the first month, 1/11 is is taken as income in the second month, ..., and 1/1 is taken as income in the 12th month.
  • Monthly withdrawals with a 5-month withdrawal cushion: Monthly withdrawals are calculated as ((balanced portfolio + cash cushion) X VPW Table percentage / 12). The withdrawal cushion matches the rate of inflation.
  • Monthly withdrawals with a 15-month withdrawal cushion: Monthly withdrawals are calculated as ((balanced portfolio + cash cushion) X VPW Table percentage / 12). The withdrawal cushion matches the rate of inflation.
  • Monthly withdrawals with a 23-month withdrawal cushion: Monthly withdrawals are calculated as ((balanced portfolio + cash cushion) X VPW Table percentage / 12). The withdrawal cushion matches the rate of inflation.


Image

We see that the red line, representing the (default) 5-month withdrawal cushion remains almost identical (with tiny fluctuations) to the black line of annual withdrawals. In contrast, both the 15-month and 23-month withdrawal cushion lines accumulate lags (of 5% and 8%, respectively) by age 87.
4nursebee wrote: Sat Aug 29, 2020 2:46 am Our situation:
Fixed income: Pension #1 now, an annuity until first of us reaches age 70, and real estate cover a lot of our expenses. In the future real estate will get paid off and that income will double, SS will kick in for both of us, 2nd pension.
Assets: Cash (non retirement acct, still invested, aka emergency fund) has grown to >7x. Roths might be 28 x. IRA/401K at least 30X. There is another annuity we need to set up to pay out at some time in the future, currently 8x, advised to set this up to pay out last.
Spouse in 1st year retirement, I've just gone part time. We did a Roth conversion, not yet to top of tax bracket, perhaps another 50K to go. My plan would be to follow VPW fully starting next year. I would add up all assets, calculate a withdrawal amount, withdraw that amount ( or a monthly equivalent?) in January. And to be clear, I would calculate withdrawal amount and subtract out the annuity payment first. Otherwise I could base withdrawal amount on non annuity funds. Withdraw for me is a bit of just mental accounting as I'd pull money from the IRA/401K portion, add it to cash portion, pay bills from that. I plan to pull from Roths last. By pulling heavy from tax deferred accounts first I'd hope to take advantage of historically low tax rates, decrease RMDs in the future, help make SS less taxed. I could just pull what we need which would be much less but think this would impact us long term. I might even convert more to a Roth this year or else pull a bit more out and add it to "cash" bucket this year to take us to top of tax bracket. I think some would argue to delay withdrawals, spend from cash bucket first, but we have been advised that our taxes will likely grow in retirement and seek to minimize this.

Questions:
What withdrawal method do you favor, the monthly method?
I'm a big fan of the VPW Accumulation And Retirement Worksheet using monthly withdrawals along with a 5-month withdrawal cushion.
4nursebee wrote: Sat Aug 29, 2020 2:46 am Does my plan sound reasonable above?
Your plan is sufficiently complex and personal to deserve its own thread (ideally with all the details suggested in this post).
4nursebee wrote: Sat Aug 29, 2020 2:46 am With a mix of assets, what amounts should be used to calculate starting portfolio and VPW? I assume either not using the annuity payout of value is good. What about cash or emergency funds and Roths?
Am I missing something? Do I need to clarify anything?
These are all good questions for the thread I suggested creating.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
LakerP
Posts: 9
Joined: Fri May 08, 2020 2:35 pm

Re: Variable Percentage Withdrawal (VPW)

Post by LakerP »

The VPW table in the wiki includes nothing for 100% stocks portfolios or anything for someone younger than 40, how can I calculate such withdrawal numbers in these situations? Also how would VPW work for someone say retiring at 30 with no social security/pension? I do not understand the purpose of keeping a social security bridge amount until reaching SS age
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

LakerP wrote: Sat Aug 29, 2020 4:50 pm The VPW table in the wiki includes nothing for 100% stocks portfolios or anything for someone younger than 40, how can I calculate such withdrawal numbers in these situations? Also how would VPW work for someone say retiring at 30 with no social security/pension? I do not understand the purpose of keeping a social security bridge amount until reaching SS age
LakerP, I encourage you start here and learn about bonds and their role in a portfolio. This thread isn't the place for that.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
LakerP
Posts: 9
Joined: Fri May 08, 2020 2:35 pm

Re: Variable Percentage Withdrawal (VPW)

Post by LakerP »

longinvest wrote: Sat Aug 29, 2020 5:17 pm
LakerP wrote: Sat Aug 29, 2020 4:50 pm The VPW table in the wiki includes nothing for 100% stocks portfolios or anything for someone younger than 40, how can I calculate such withdrawal numbers in these situations? Also how would VPW work for someone say retiring at 30 with no social security/pension? I do not understand the purpose of keeping a social security bridge amount until reaching SS age
LakerP, I encourage you start here and learn about bonds and their role in a portfolio. This thread isn't the place for that.
I understand the role of bonds in a portfolio and have no interest in including them, I didn't come here to debate portfolio theory. Is this withdrawal method unviable without a SS/pension? Is it only valid for someone that is at least 40 years old?
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

LakerP wrote: Sat Aug 29, 2020 5:20 pm
longinvest wrote: Sat Aug 29, 2020 5:17 pm
LakerP wrote: Sat Aug 29, 2020 4:50 pm The VPW table in the wiki includes nothing for 100% stocks portfolios or anything for someone younger than 40, how can I calculate such withdrawal numbers in these situations? Also how would VPW work for someone say retiring at 30 with no social security/pension? I do not understand the purpose of keeping a social security bridge amount until reaching SS age
LakerP, I encourage you start here and learn about bonds and their role in a portfolio. This thread isn't the place for that.
I understand the role of bonds in a portfolio and have no interest in including them, I didn't come here to debate portfolio theory. Is this withdrawal method unviable without a SS/pension? Is it only valid for someone that is at least 40 years old?
Dear LakerP,

Have you taken the time to read the detailed post I wrote for you a few days ago? I didn't see any reply from you.

Best regards,

longinvest
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
LadyGeek
Site Admin
Posts: 68645
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Variable Percentage Withdrawal (VPW)

Post by LadyGeek »

I removed an off-topic post. As a reminder, see: General Etiquette
At all times we must conduct ourselves in a respectful manner to other posters. Attacks on individuals, insults, name calling, trolling, baiting or other attempts to sow dissension are not acceptable.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.
Brit
Posts: 70
Joined: Wed Oct 10, 2012 6:33 am

Re: Variable Percentage Withdrawal (VPW)

Post by Brit »

+1 LadyGeek, longinvest freely, generously, patiently and respectfully shares his knowledge and answers questions to help other people grow and learn. The same courtesy should be returned.
rama13
Posts: 46
Joined: Sun May 15, 2011 2:09 am

Re: Variable Percentage Withdrawal (VPW)

Post by rama13 »

longinvest,

Have you considered adding functionality to the spreadsheet to do the calculations for the withdrawal cushion adjustment exemplified below?
longinvest wrote: Sat Sep 05, 2020 11:41 amThe VPW Worksheet suggests to take a $5,675 withdrawal. The retiree withdraws the suggested amount.

After making the withdrawal, the retiree calculates an adjustment using the withdrawal cushion balance:
  • Adjusted withdrawal amount: (($5,675 + $26,868.87) / 6) = $5,424
  • Adjustment: ($5,424 - $5,675) = -$251
It seems that instead of having a single entry for total portfolio amount, there could be two entries--portfolio balance (excluding the withdrawal cushion) and withdrawal cushion. This would mirror your inputs here:
longinvest wrote: Sat Sep 05, 2020 11:41 am As of August 31, 2020, the retiree's portfolio is composed of:
  • Vanguard LifeStrategy Moderate Growth Fund (VSMGX): $1,009,096.40
  • Withdrawal cushion (at Ally Bank): $26,868.87
Total Portfolio Balance: $1,035,965.27

...

We update the Portfolio Balance cell in the Retirement sheet of the VPW Accumulation And Retirement Worksheet. No other entry needs updating.
The output could then be the withdrawal cushion adjustment (or in other words, how much of the withdrawal should I put/take from the cushion vs. the rest of the portfolio).
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

rama13 wrote: Sun Sep 06, 2020 2:21 pm Have you considered adding functionality to the spreadsheet to do the calculations for the withdrawal cushion adjustment exemplified below?
[...]
Rama13, thanks for the suggestion, but in the VPW forward test thread, transactions and transfers neatly happen all at once at the end of the last day of the month.

Things are likely to differ in real life. The sell transaction might happen earlier during the month. Once the transaction settles, money might get transferred to the withdrawal cushion. On the first day of the following month, once interest has been paid into the savings account, one sixth of the withdrawal cushion might get transferred to the spending bank account. There are many reasonable variations like that.

The spreadsheet is free and open. It's easy to modify and adapt to one's specific personal financial plan.

I prefer to keep the wiki's spreadsheet simple. Sometimes less is more.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
JasonFIRE
Posts: 6
Joined: Fri Sep 18, 2020 9:30 pm

Re: Variable Percentage Withdrawal (VPW)

Post by JasonFIRE »

Dear longinvest,

Thank you very much for all of your efforts on VPW. It is very impressive. I am close to wanting to adopt it for my situation, but I have some nagging questions.

Many times I have read your description of the growth trends and the concept of the WAG, and it makes sense. I have also read this thread starting from your post on page 17. So, there is no need to reexplain that. My questions about the growth trends and WAG are mainly HOW it is used in the calculations. I would be grateful if you would be willing to elaborate on that. Would it have a significant effect if it were changed?

My other main question is about the table. I would also be grateful if you could elaborate on HOW the table is created/calculated. I cannot seem to find any details on that, other than perhaps digging through this thread on pages 1-17.

Basically, I love the concept of VPW, but my mind would be more at peace if I had more of an understanding of the above, instead of just blindly accepting some guy's numbers on the internet. In mean that respectfully; clearly you and the community have done amazing work on this.

Many thanks in advance!
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

JasonFIRE,
JasonFIRE wrote: Fri Sep 18, 2020 9:35 pm Dear longinvest,

Thank you very much for all of your efforts on VPW. It is very impressive. I am close to wanting to adopt it for my situation, but I have some nagging questions.

Many times I have read your description of the growth trends and the concept of the WAG, and it makes sense. I have also read this thread starting from your post on page 17. So, there is no need to reexplain that. My questions about the growth trends and WAG are mainly HOW it is used in the calculations. I would be grateful if you would be willing to elaborate on that. Would it have a significant effect if it were changed?
Upon reaching age 60 in 1981, two hypothetical investors retire. Each has a $1,000,000 60/40 stock/bond portfolio (stocks divided 60/40 U.S./international) and no pension. The first retiree doesn't know future returns and uses VPW with its ((1.9% bonds X 40%) + (5.0% stocks X 60%)) = 3.76% growth trend WAG. The second retiree has perfect foresight and knows exactly how much a 60/40 stock/bond portfolio will return over the next 39 years and uses this Future39 annualized return as growth trend. (The return of the 40th year, 2020, isn't needed as the portfolio gets depleted at the start of the 40th year).

Here's how their withdrawals compare in inflation-adjusted dollars:

Image

The "Future39 Withdrawal" line represents withdrawals using the 6.36% 39-year annualized real return of the portfolio from 1981 to 2019 (inclusively) as growth trend. Despite the large difference between 6.36% and 3.76%, yearly ups and downs remain quite similar. What using the exact future long-term return as growth trend did was only to equalize the first and the last withdrawals.

Here's a similar chart for 1937, a challenging retirement start year with a 3.38% future 39-year annualized real return:

Image

As we can see, if the WAG turns out lower than long-term portfolio returns, the VPW line tilts to the left relative to the Future39 line, and if the WAG turns out higher than long-term portfolio returns, the VPW line tilts to the right relative to the Future39 line.

For perspective, using VPW backtesting spreadsheet data, the highest 39-year annualized real return of the selected 60/40 portfolio (where stocks were U.S.-only until 1971 and 60/40 U.S./international starting in 1972) was 7.33% in 1921 and the lowest 39-year annualized real return was 3.24% in 1881. Other countries might not have been as lucky.
JasonFIRE wrote: Fri Sep 18, 2020 9:35 pm My other main question is about the table. I would also be grateful if you could elaborate on HOW the table is created/calculated. I cannot seem to find any details on that, other than perhaps digging through this thread on pages 1-17.

Basically, I love the concept of VPW, but my mind would be more at peace if I had more of an understanding of the above, instead of just blindly accepting some guy's numbers on the internet. In mean that respectfully; clearly you and the community have done amazing work on this.

Many thanks in advance!
Here's a simple way to reconstruct the 60/40 stocks/bonds column of the VPW Table with simple arithmetic operations.

Let's assume that we want to withdraw $1 per year from our portfolio on our birthday, during retirement, until age 99 when our portfolio gets depleted. We also assume that the annual return of the portfolio is exactly 3.76% each year of retirement.

Upon reaching age 99, we make our last $1 withdrawal and deplete our portfolio. Obviously our portfolio balance, before withdrawal, was exactly $1. We withdrew $1 / $1 = 100.0% of our portfolio at age 99.

This implies that at age 98, just after making our $1 withdrawal, we had $1 / (1 + 3.76%) = $0.963762528 portfolio balance. (Check: If we let $0.963762528 grow by 3.76% during our 98th year, we get a $1 portfolio balance at age 99 before withdrawal). Before making our withdrawal at age 98 we necessarily had $1.963762528. We withdrew $1 / $1.963762528 = 50.9% (rounded) of our portfolio at age 98.

We repeat the calculation for age 97 and get $1 + ($1.963762528 / (1 + 3.76%)) = $2.892600741, implying a $1 / $2.892600741 = 34.6% withdrawal.

We repeat the calculation for age 96 and get $1 + ($2.892600741 / (1 + 3.76%)) = $3.787780205, implying a $1 / $3.787780205 = 26.4% withdrawal.

And so on for ages 96, 95, ..., and 40.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
User avatar
siamond
Posts: 5724
Joined: Mon May 28, 2012 5:50 am

Re: Variable Percentage Withdrawal (VPW)

Post by siamond »

JasonFIRE wrote: Fri Sep 18, 2020 9:35 pmMany times I have read your description of the growth trends and the concept of the WAG, and it makes sense. I have also read this thread starting from your post on page 17. So, there is no need to reexplain that. My questions about the growth trends and WAG are mainly HOW it is used in the calculations. I would be grateful if you would be willing to elaborate on that. Would it have a significant effect if it were changed?

My other main question is about the table. I would also be grateful if you could elaborate on HOW the table is created/calculated. I cannot seem to find any details on that, other than perhaps digging through this thread on pages 1-17.
Let me answer your question in another way. The annual VPW withdrawal (in absolute $) can be directly computed with a simple PMT() function, e.g. in a spreadsheet: annual-withdrawal = -PMT(rate, nper, pv, fv, type) where:

- rate is the WAG derived from your stocks/bonds asset allocation
- nper is the number of remaining years
- pv is the present value of your portfolio
- fv (final value) is equal to zero (aiming to deplete the portfolio after 'nper' years)
- type is equal to 1 to indicate a withdrawal at the beginning of the year

And yes, as you probably suspected, the output of the formula year over year is quite sensitive to the 'rate' parameter. Using longinvest's WAG is just one way of proceeding among others.
Normchad
Posts: 1785
Joined: Thu Mar 03, 2011 7:20 am

Re: Variable Percentage Withdrawal (VPW)

Post by Normchad »

@longinvest Thank you so much for all the work you’ve put into this spreadsheet and the continuing explanations of it. I’ve read a lot about it over the last few years, but only started looking at it seriously about a week ago. It’s very nice, very simple, and the results are quite reassuring.

Best of all, this just might be something my widow could manage/maintain if I predecease her.
Topic Author
longinvest
Posts: 4534
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Post by longinvest »

The use of a constant (unchanging) growth trend is fundamental to VPW. It avoids the "selling less high, selling more low" behavior.

I'll continue illustrating examples with the same 60/40 stock/bond portfolio as my previous post using data from the VPW backtesting spreadsheet.

Here's an example of why I think that using a 10-year forward prediction as growth trend WAG is a bad idea:

Upon reaching age 60 in 1972, two hypothetical investors retire. Each has a $1,000,000 60/40 stock/bond portfolio (stocks divided 60/40 U.S./international) and no pension. The first retiree doesn't know future returns and uses VPW with its ((1.9% bonds X 40%) + (5.0% stocks X 60%)) = 3.76% growth trend wild ass guess (WAG). The second retiree has a perfect foresight about the forward 10 year return and knows that the portfolio will return exactly -1.16% real annualized from 1972 to 1981 (inclusively) and uses it as growth trend to derive a VPW table for retirement. Here's the outcome:

Image

Perfect knowledge about future 10-year returns isn't helpful; using it as constant growth trend is actually harmful as it leads the retiree to severely underspend at the start of retirement with an initial $19,720 withdrawal representing less than 2% of the portfolio.
siamond wrote: Sun Sep 20, 2020 4:14 am Using longinvest's WAG is just one way of proceeding among others.
Siamond is indirectly referring to his PMT withdrawal method which consists of using the PMT excel function combined with a 10-year return prediction to derive a withdrawal percentage for the current year.

If instead of using the prefect 10-year forward prediction as a constant growth trend WAG to derive a VPW Table, in the above example, the second retiree had used a perfect 10-year return prediction every year with Siamond's PMT withdrawal method, the outcome would have been:

Image

Using a perfect forward 10-year prediction leads to taking a $19,720 withdrawal in 1972. Then, as the portfolio is losing -38% of its value from 1973 to 1975, withdrawals are increased by 64% from $23,699 to $38,935. In 1977, the withdrawal reaches its maximum of $58,121, almost tripling the initial withdrawal amount in inflation-adjusted terms during a tumultuous 5 years period.

It's difficult to understand why withdrawals then drop by -26% from 1977 to 1986, and again drop by -31% from 1990 to 1993, during a raging bull market. It's much easier to understand why withdrawals drop by -24% in 2009; it's because the portfolio's time-weighted return was -22% in 2008 and, with only 2 withdrawals left, playing with PMT's rate can't hide portfolio losses.

The cost of trying to "smooth" withdrawals with perfect return predictions was high. It lead to getting -22% less in inflation-adjusted withdrawals with PMT than with VPW. (Average PMT withdrawal: $44,658, average VPW withdrawal: $56,667). "Sell less high, sell more low" is a form of "buy high, sell low", one the the worst possible investing approaches. As I said, the use of a constant (unchanging) growth trend is fundamental to VPW; it's meant, among other things, to avoid such harmful investing behavior. It's also humanly natural to reduce withdrawals in bad times, when the portfolio is sustaining losses, and increase them in good times, when the portfolio is gaining in value.

This thread isn't about Siamond's PMT method and its use of return prediction metrics. It's about VPW.
Last edited by longinvest on Thu Oct 22, 2020 10:17 pm, edited 1 time in total.
Bogleheads investment philosophy | One-ETF global balanced index portfolio | VPW
Post Reply