Variable Percentage Withdrawal (VPW)

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat Jun 13, 2015 3:02 pm

mouth wrote:1) Pension - not all pensions are NOCOLA


True, but if a pension is COLA, you can simply add it to the Social Security amount.

Effectively, the entries labeled as "Social Security" and "Pension" are, actually, "COLA pensions" and "NOCOLA pensions" entries. It just seems more intuitive for new users, to use the current labels.

mouth wrote:2) Social Security - It seems to start at StartAge vice 62/67/70 ... problem is I plan on retiring at 43.


This is so young that, personally, I would simply not count the Social Security pension for two reasons:
  1. I would not accumulate much benefits, not contributing between 43 and Social Security's Full Retirement Age.
  2. (Assuming a delayed-to-70 SS pension), 27 years is a very long period of time until the first SS payment. I would simply revisit the situation around 60, possibly extracting a bridging amount from the portfolio to put into CDs, and starting to consider the expected SS@70. (See Delay Social Security to age 70 and Spend more money at 62).

mouth wrote:3) The same can be said for pensions. Mine actually has started already, but others might want to input a different start age than StartAge if they plan to retire earlier than their pension starts.


Same answer as 2).

mouth wrote:Not huge deals since backtesting is just a nice to have feature compared to the main VPW tab. Thanks for all the hard work!!!


You're welcome. I'm glad this tool is useful to you.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

Christine600
Posts: 18
Joined: Sun May 03, 2015 12:09 am

Re: Variable Percentage Withdrawal (VPW)

Postby Christine600 » Mon Jun 15, 2015 4:21 pm

Hi and thank you to Longinvest for all this work!


I retired last year and have started using VPW for this year. That is I entered last years figures too with 0,- withdrawn.

I'm not very familiar with spreadsheets nor financial mathematics. So I have not been able to find the ansver to one question.

Does it matter which currency you use? I live in Europe and use NOK. But so far I have just entered the correct numbers for me ignoring the $ sign. My idea is if all numbers and formulas use percentages then it should not matter if my 1000 is different from $1000. All calculations would be correct anyway.

This would be similar to someone using the spreadsheet and entering numbers in thousands - ie entering $10,000 if they in reality they have $ 10,000,000. If they then got a Suggested Withdrawa of $400 then they could withdraw $400 000.

Am I making any sense? :D

BahamaMan
Posts: 896
Joined: Wed Oct 01, 2014 5:52 pm

Re: Variable Percentage Withdrawal (VPW)

Postby BahamaMan » Mon Jun 15, 2015 4:40 pm

Christine600 wrote:Does it matter which currency you use? I live in Europe and use NOK. But so far I have just entered the correct numbers for me ignoring the $ sign. My idea is if all numbers and formulas use percentages then it should not matter if my 1000 is different from $1000. All calculations would be correct anyway.

This would be similar to someone using the spreadsheet and entering numbers in thousands - ie entering $10,000 if they in reality they have $ 10,000,000. If they then got a Suggested Withdrawa of $400 then they could withdraw $400 000.

Am I making any sense? :D


Not Longinvest, but I will answer your question anyway.

Yes, you are making perfect sense! .... Yes, the percentages are what it's all about, so it does not matter what currency you are calculating in. As long as you are taking the percentage * remaining portfolio balance, you will calculate the annual spending level in any currency you desire.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Mon Jun 15, 2015 4:57 pm

Christine600 wrote:Hi and thank you to Longinvest for all this work!


You're welcome.

Christine600 wrote:Does it matter which currency you use? I live in Europe and use NOK. But so far I have just entered the correct numbers for me ignoring the $ sign. My idea is if all numbers and formulas use percentages then it should not matter if my 1000 is different from $1000. All calculations would be correct anyway.


You are right. I agree with BahamaMan.

Christine600 wrote:Am I making any sense? :D


Yes you are.

It would add additional complexity to provide a choice of currency symbols in the spreadsheet. I feel that the spreadsheet is already pretty complex. I want it to remain simple enough for people to understand how it works, and see that there are no hidden magic formulas. Just simple mathematics.

Note that the spreadsheet is modifiable. It is quite easy to simply go ahead and to change the currency symbol in one's personal copy so that it looks right.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
LadyGeek
Site Admin
Posts: 38165
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Variable Percentage Withdrawal (VPW)

Postby LadyGeek » Tue Sep 22, 2015 7:45 pm

digarei has supplied feedback to the wiki's reader feedback tool. I added the comment to the wiki's Discussion tab:

The VPW spreadsheet downloaded from Google drive does not work in all versions of Excel... it will not load, for example, into Office Excel for iPad (v1.13) [running under iOS 8.4.1 on iPad Air 2]. However, it does function reasonably well using Apple's spreadsheet Numbers (v2.5.5) on the iPad. I downloaded several other spreadsheets (Calc XL, iSpreadsheet) to try but none of them retained what appears to be named ranges and other features, so dead in the water. Bottom Line: MS Excel on iPad (NO), Apple Numbers (YES--the charts weren't pretty). - digarei 9/21/2015
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.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Tue Sep 22, 2015 8:04 pm

LadyGeek wrote:digarei has supplied feedback to the wiki's reader feedback tool. I added the comment to the wiki's Discussion tab:

The VPW spreadsheet downloaded from Google drive does not work in all versions of Excel... it will not load, for example, into Office Excel for iPad (v1.13) [running under iOS 8.4.1 on iPad Air 2]. However, it does function reasonably well using Apple's spreadsheet Numbers (v2.5.5) on the iPad. I downloaded several other spreadsheets (Calc XL, iSpreadsheet) to try but none of them retained what appears to be named ranges and other features, so dead in the water. Bottom Line: MS Excel on iPad (NO), Apple Numbers (YES--the charts weren't pretty). - digarei 9/21/2015


If digarei is reading this thread: Please try the DropBox download. Theoretically, it shouldn't make a difference, but in practice, sometimes it does.

I guess that it is time to replicate the work I did with the Bogleheads Returns spreadsheet: I developed three distinct copies in free Office online (Excel), Google Docs (Sheets), and LibreOffice (Calc). I've heard of no compatibility complaints since then. (Yay!) I think that some people really like the online Google Docs (Sheets) version.

It's just that VPW is a much bigger spreadsheet and will require a lot more work.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Mon Sep 28, 2015 9:52 am

Readers of this thread might be interested by two posts I have made on another thread ("Cost to Fund Retirement"):
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Fri Jan 08, 2016 11:18 am

I am copying here a question asked in another thread:
Bustoff wrote:
longinvest wrote:
So, my preferred method is VPW which adapts withdrawals to the retiree's retirement horizon, asset allocation, and portfolio returns during retirement. VPW is not only mathematically guaranteed to never deplete the portfolio prematurely, it will also let the retiree to enjoy his savings while alive.

I have tried using the spreadsheet but have difficulty understanding it despite consulting the thread. (I have no experience using spreadsheets.)
You put considerable time and effort into the VPW calculator and I would really like to use it. Any suggestions?
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
BlueEars
Posts: 3471
Joined: Sat Mar 10, 2007 12:15 am
Location: West Coast

Re: Variable Percentage Withdrawal (VPW)

Postby BlueEars » Fri Jan 08, 2016 11:53 am

A few suggestions for Bustoff assuming you've got the VPW spreadsheet on your computer.
1) Read the Instructions sheet (tab named this)
2) Put in your data on the VPW sheet as per instructions
3) Get familiar with the Backtesting sheet. It has similar inputs as the VPW sheet but you can set the Start Year and see how the portfolio does in inflation adjusted terms. Some bad start year sequences were 1929 and 1968. The chart to the right of the data shows all start years and minimum withdrawals (worst case).
4) I just discovered this recently. If you go to the Table sheet you can override the start percentage to anything you choose (select Yes in cell E16, set percentage in E17). Then you can see the results in the Backtesting sheet. Also you can now set the Last Withdrawal Age to a large number like 1000. Now you get the backtesting for a constant withdrawal amount. Personally I'd like to see these controls in the Backtesting sheet but it's OK the way it is although a bit obscure -- or maybe there is a better way I don't know about.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Fri Jan 08, 2016 12:03 pm

BlueEars wrote:A few suggestions for Bustoff assuming you've got the VPW spreadsheet on your computer.
1) Read the Instructions sheet (tab named this)
2) Put in your data on the VPW sheet as per instructions
3) Get familiar with the Backtesting sheet. It has similar inputs as the VPW sheet but you can set the Start Year and see how the portfolio does in inflation adjusted terms. Some bad start year sequences were 1929 and 1968. The chart to the right of the data shows all start years and minimum withdrawals (worst case).
4) I just discovered this recently. If you go to the Table sheet you can override the start percentage to anything you choose (select Yes in cell E16, set percentage in E17). Then you can see the results in the Backtesting sheet. Also you can now set the Last Withdrawal Age to a large number like 1000. Now you get the backtesting for a constant withdrawal amount. Personally I'd like to see these controls in the Backtesting sheet but it's OK the way it is although a bit obscure -- or maybe there is a better way I don't know about.

BlueEars,

Thanks for the help. I had no idea where to start explaining things.

Just a few minor details:

To compare VPW with the flawed SWR model, there's already a Comparison sheet; there's no need to tinker with the Table sheet. (The Table sheet is only provided for advanced uses). There's a "Withdrawal Percentage" parameter on the top of the Comparison sheet to compare other rates than the traditional 4%.

Setting the Last Withdrawal Age to 1000 will simulate "constant percentage", which is different from SWR ("constant dollar").
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
Bustoff
Posts: 1979
Joined: Sat Mar 03, 2012 6:45 pm

Re: Variable Percentage Withdrawal (VPW)

Postby Bustoff » Fri Jan 08, 2016 12:44 pm

Thanks BlueEars and longinvest. I appreciate your suggestions.
I'm going to spend some time on this today and this weekend.

User avatar
BlueEars
Posts: 3471
Joined: Sat Mar 10, 2007 12:15 am
Location: West Coast

Re: Variable Percentage Withdrawal (VPW)

Postby BlueEars » Fri Jan 08, 2016 1:14 pm

I never even looked at the Comparison sheet, thanks Longinvest.

I think a lot of retirees struggle with the desire to see their portfolio maintain altitude and yet desire to not leave too much on the table. I have used VPW as my upper limit on spending. Last year we spent 3.5% but could have spend 4.5% according to VPW. We got everything we really wanted that year (discretionary spending) and the markets weren't that kind last year so it seems to have worked out. I'll mentally remember that 1.0% holdout when and if the market tanks and spending should be decreased (according to VPW). But then again, we never know if the forward sequence might not be one historically even worse then the past. :?

If you are putting out a revision Longinvest, I'd suggest you mention the Comparison sheet in the Instructions sheet. You might have a short Beginners section, and an Advanced section. Shouldn't be much work compared to all the actual (more fun and harder) analytical work. Full disclosure: I never even read the Instructions sheet :oops: .

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Fri Jan 08, 2016 1:28 pm

BlueEars wrote:I think a lot of retirees struggle with the desire to see their portfolio maintain altitude and yet desire to not leave too much on the table. I have used VPW as my upper limit on spending. Last year we spent 3.5% but could have spend 4.5% according to VPW. We got everything we really wanted that year (discretionary spending) and the markets weren't that kind last year so it seems to have worked out. I'll mentally remember that 1.0% holdout when and if the market tanks and spending should be decreased (according to VPW). But then again, we never know if the forward sequence might not be one historically even worse then the past. :?

Personally, I would take the full VPW withdrawal and transfer any excess money into a high-interest savings account, instead. Otherwise, the unwithdrawn amount will melt along with the rest of my portfolio during a market downturn, if I leave it invested.

I actually learned about this when Cut-Throat was exchanging with me about the VPW spreadsheet.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
BlueEars
Posts: 3471
Joined: Sat Mar 10, 2007 12:15 am
Location: West Coast

Re: Variable Percentage Withdrawal (VPW)

Postby BlueEars » Fri Jan 08, 2016 1:45 pm

Or maybe short term bonds for the unspent bookkeeping, such as vfsux or vfisx.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Fri Jan 08, 2016 1:48 pm

BlueEars wrote:Or maybe short term bonds for the unspent bookkeeping, such as vfsux or vfisx.

Yes, that's the spirit.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

BahamaMan
Posts: 896
Joined: Wed Oct 01, 2014 5:52 pm

Re: Variable Percentage Withdrawal (VPW)

Postby BahamaMan » Fri Jan 08, 2016 4:06 pm

longinvest wrote:Personally, I would take the full VPW withdrawal and transfer any excess money into a high-interest savings account, instead. Otherwise, the unwithdrawn amount will melt along with the rest of my portfolio during a market downturn, if I leave it invested.

I actually learned about this when Cut-Throat was exchanging with me about the VPW spreadsheet.


+1 .....That is exactly what I do. I always withdraw the Full VPW amount from my Investment Portfolio into my Cash spending account. This cash is not part of my asset allocation, and will never be invested in anything other than Cash.

The VPW withdrawal amount gets you to sell your Investments when the market is high and pull back when it is low. It is a discipline just like dollar cost averaging is. It forces you to sell at higher prices and Less with market declines. It takes the 'emotion' out of withdrawals. The Withdrawal Amount drives the Budget, not the other way around!

User avatar
Bustoff
Posts: 1979
Joined: Sat Mar 03, 2012 6:45 pm

Re: Variable Percentage Withdrawal (VPW)

Postby Bustoff » Sat Jan 09, 2016 1:06 pm

I get a Microsoft Office Excel Security Alert when I try to open the download.
Tried both the Dropbox and the Google Drive download links but get this:

Image

Should I ignore the alert?

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat Jan 09, 2016 1:29 pm

Bustoff wrote:I get a Microsoft Office Excel Security Alert when I try to open the download.
Tried both the Dropbox and the Google Drive download links but get this:
[...]
Should I ignore the alert?


You may safely ignore the alert. Please read the note in the Wiki page:
https://www.bogleheads.org/wiki/Variabl ... soft_Excel
Microsoft Excel may raise an Office File Validation security error. This is because the spreadsheet was not built using Microsoft software; it was built using OpenOffice Calc and saved as Microsoft Excel format. Consequently, Excel raises a warning to the user. This error can be safely ignored.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat Jan 09, 2016 1:30 pm

This morning, I have added two new sections to the wiki: Variable percentage withdrawal page:

  • How to use the VPW spreadsheet during retirement
  • How to use VPW without a spreadsheet program during retirement
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
Bustoff
Posts: 1979
Joined: Sat Mar 03, 2012 6:45 pm

Re: Variable Percentage Withdrawal (VPW)

Postby Bustoff » Sat Jan 09, 2016 1:31 pm

longinvest wrote:This morning, I have added two new sections to the wiki: Variable percentage withdrawal page:

  • How to use the VPW spreadsheet during retirement
  • How to use VPW without a spreadsheet program during retirement



Wow ... Thank you so much for your contribution!

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat Jan 09, 2016 1:37 pm

In another thread, Don asked two questions.

Don46 wrote:I have two questions: 1) the "portfolio" amount should include all my (and my wife's) retirement accounts and taxable savings accounts, correct?
2) the variation in the withdrawals is determined by what? Market performance for the previous year?

Many thanks for all this help from everyone.
--Don
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat Jan 09, 2016 1:47 pm

Don46 wrote:1) the "portfolio" amount should include all my (and my wife's) retirement accounts and taxable savings accounts, correct?


Correct, except that you should exclude from the calculation any money set aside for other purposes, such as:
  • Money set aside to bridge retirement and the start of Social Security payments.
  • Emergency fund.
  • Any money withdrawn using VPW. Any withdrawn money that is not needed should be kept in cash or near cash (money-market, savings account, etc.) and never be reinvested. (We certainly don't want this money to melt with the portfolio in an upcoming bear market).

You might want to read the new "How to use..." Section that I have added this morning to the VPW wiki page.

Don46 wrote:2) the variation in the withdrawals is determined by what? Market performance for the previous year?


Yes, mostly. It's a combination of age (has increased, affecting the withdrawal percentage) and previous year market returns (affecting the size of the portfolio on which the percentage is applied).
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
LadyGeek
Site Admin
Posts: 38165
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Variable Percentage Withdrawal (VPW)

Postby LadyGeek » Sat Jan 09, 2016 2:27 pm

Is there anything that needs to be updated to your Canadian version, Variable percentage withdrawal - finiki, the Canadian financial wiki? I recently added this reference in the "See also" section of the wiki page (and vice versa).

The corresponding support thread is here: Variable Percentage Withdrawal (VPW) for Canadians, in the Financial Wisdom Forum.
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.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat Jan 09, 2016 2:33 pm

LadyGeek wrote:Is there anything that needs to be updated to your Canadian version, Variable percentage withdrawal - finiki, the Canadian financial wiki? I recently added this reference in the "See also" section of the wiki page (and vice versa).

LadyGeek,

Thanks for the help offer. One possible thing would be to copy the Bogleheads wiki page into Finiki, but to replace "Social Security" with "Canada/Quebec Pension Plan (CPP/QPP) and Old Age Security (OAS)".

P.S. There might be some English syntax problems in my texts. Feel free to fix any that you see.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
LadyGeek
Site Admin
Posts: 38165
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Variable Percentage Withdrawal (VPW)

Postby LadyGeek » Sat Jan 09, 2016 3:24 pm

May I suggest inserting a revision date in your spreadsheet?

The changes tab has 01-28-2015 as the last date. I recommend adding an entry for your revision of 01-09-2016.

The wiki (and finiki) pages are showing a version date of January 28, 2015 - which is not accurate.


Update: See below. I misinterpreted what longinvest said.
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.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat Jan 09, 2016 3:29 pm

Right! OK, I'll go and clean things up later today.

My longer-term objective is to include some suggestions you made on the FWF forum and develop software-specific versions (like wiith the personal returns spreadsheet).
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
artthomp
Posts: 315
Joined: Tue Feb 20, 2007 4:44 pm
Location: Ballwin Missouri

Re: Variable Percentage Withdrawal (VPW)

Postby artthomp » Sat Jan 09, 2016 3:39 pm

I'm 75 years old and I have a pension and my wife and I are taking Social Security so I have been able to rebalance to 2% cash from the portfolio every year so far. 2002 and 2008 are the only years in which the portfolio did not reach an all time high sometime during the year. I do have to take RMDS distributions and I deduct my Federal and State taxes from the distribution and than reinvest any excess rebalancing to 2% cash for any additional expenses.
Art

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat Jan 09, 2016 7:26 pm

LadyGeek,

Thanks for the Wiki page structure improvements.

About your previous post:

LadyGeek wrote:May I suggest inserting a revision date in your spreadsheet?

The changes tab has 01-28-2015 as the last date. I recommend adding an entry for your revision of 01-09-2016.

The wiki (and finiki) pages are showing a version date of January 28, 2015 - which is not accurate.


In both Dropox and GoogleDrive, as well as on my personal computer, the file has a last modification date of January 28, 2015. Are you expecting me to upload a new version today?

Today, I have not modified the spreadsheet; I have only improved the wiki page. Yes, I am overdue for integrating a suggestion you made on the FWF forums, but I am waiting for CPI-U (US) and CPI (Canada) data to update the data sets and implement some minor improvements.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
LadyGeek
Site Admin
Posts: 38165
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Variable Percentage Withdrawal (VPW)

Postby LadyGeek » Sat Jan 09, 2016 7:36 pm

longinvest wrote:Are you expecting me to upload a new version today?

No, this is my mistake. I had assumed that you had updated the spreadsheet along with the wiki page. I fixed my previous post.
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.

User avatar
Leif
Posts: 2086
Joined: Wed Sep 19, 2007 4:15 pm

Re: Variable Percentage Withdrawal (VPW)

Postby Leif » Wed Feb 10, 2016 4:11 pm

I'm confused when I try to compare the withdrawal rate table on the Wiki to the withdrawal rate in the spreadsheet.

The Wiki shows for age 63 and an allocation 50/50 with WR is 4.7%.

When I enter 25% domestic, 25% international (50% total) and 50% bonds the spreadsheet reports 5.2%

Is there some additional factor used, involving domestic/international split, that is not represented in the Wiki table?
Investors should diversify across many asset-classes so that whatever happens, we will not have all our investments in underperforming asset classes and thereby fail to meet our goals-Taylor Larimore

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Wed Feb 10, 2016 4:22 pm

Leif wrote:I'm confused when I try to compare the withdrawal rate table on the Wiki to the withdrawal rate in the spreadsheet.

The Wiki shows for age 63 and an allocation 50/50 with WR is 4.7%.

When I enter 25% domestic, 25% international (50% total) and 50% bonds the spreadsheet reports 5.2%

Is there some additional factor used, involving domestic/international split, that is not represented in the Wiki table?

Leif,

In my copy of the spreadsheet, I get 4.7% at 63 with a 50/50 portfolio and last withdrawal age set to 99 (depletion years: 37).

The withdrawal percentage is only affected by the bond/stock ratio; it is not affected by the ratio of domestic/international stocks.

Maybe the last withdrawal age was inadvertently changed? The last withdrawal age should not be set to the "expected" longevity of the retiree; it should be set conservatively to an age he is unlikely to reach.

The other possibility is that the default values, on the "Table" sheet were changed. If that's the case, the easiest would be to simply download a clean copy of the spreadsheet.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
Leif
Posts: 2086
Joined: Wed Sep 19, 2007 4:15 pm

Re: Variable Percentage Withdrawal (VPW)

Postby Leif » Wed Feb 10, 2016 5:14 pm

You are right. I had my last withdrawal age set to 93. Thanks.
Investors should diversify across many asset-classes so that whatever happens, we will not have all our investments in underperforming asset classes and thereby fail to meet our goals-Taylor Larimore

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Wed Feb 10, 2016 5:26 pm

Leif wrote:You are right. I had my last withdrawal age set to 93. Thanks.

That looks low to me. Personally, I wouldn't reduce it below 99. Actually, I would start increasing it above 100 in my 80s, if my health is still good.
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

User avatar
BlueEars
Posts: 3471
Joined: Sat Mar 10, 2007 12:15 am
Location: West Coast

Re: Variable Percentage Withdrawal (VPW)

Postby BlueEars » Wed Feb 10, 2016 5:44 pm

There will be various considerations in setting the Last Withdrawal Age. Ours is set to 110. This is primarily to leave us a big cushion in our ancient years. Since we get a good spending number even in a bad starting retirement year sequence (like starting in 1968), I'm comfortable with this. No need to get more aggressive. Of course, the AA also comes into this picture and if bonds start paying a decent real rate I might revisit this.

User avatar
LadyGeek
Site Admin
Posts: 38165
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Variable Percentage Withdrawal (VPW)

Postby LadyGeek » Wed Feb 10, 2016 8:29 pm

The bottom of every wiki page has a form where readers can supply feedback without logging in. We just received two comments:

No suggestion. Only thank you.

I downloaded the excel spreadsheet from Dropbox and it loads and runs fine using Apple Numbers (v3.6.1)application aside from two inconsequential initial warning messages.
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.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Wed Feb 10, 2016 11:28 pm

In another thread, I stated that VPW eliminates sequence of return risk. It think that this bears repeating in this thread.

VPW eliminates sequence of return risk. As a proof, if you permute the returns of the first N years, the withdrawal will remain unchanged in year N+1. This is trivially true, because making a VPW withdrawal is simply making a multiplication, as is applying an annual return, and multiplication is commutative. Reordering the ( (2 X N) + 1) multiplications which determine the withdrawal in year N+1 will not change the result of the multiplication.

This property is not true with the SWR method (constant-dollar withdrawal). In that case, the withdrawal in year N+1 is really dependent on the exact sequence of the previous N returns; which is why we say that SWR is vulnerable to sequence of return risk.

What is left is market risk. If the market goes down, withdrawals will go down with it. VPW does not eliminate market risk. That's why it is a good idea to apply VPW on a portfolio which contains a fair amount of bonds, and to combine it with inflation-adjusted base income (like Social Security).
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Mon Feb 15, 2016 12:24 pm

I have uploaded a new version of the VPW spreadsheet.

As usual, you can find the download links on the wiki: Variable percentage withdrawal page.

Changes:
  • Added 2015 return data.
  • Use improved historical bond return approximations for pre-1972 data. (See: viewtopic.php?f=10&t=150025).
  • Started directly collecting Canadian historical returns and inflation data (for 2015 and later) based on NAV returns of Vanguard Canada ETFs and Statistics Canada CPI numbers.

As of this version, nothing was really changed in the presentation. Improvements will go in future versions.

As usual comments are welcome.

Enjoy!
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

2Birds1Stone
Posts: 315
Joined: Wed Nov 19, 2014 10:33 pm
Location: New York

Re: Variable Percentage Withdrawal (VPW)

Postby 2Birds1Stone » Fri Feb 26, 2016 2:52 pm

Thank you for the outstanding work Longinvest!

AlohaJoe
Posts: 2073
Joined: Mon Nov 26, 2007 2:00 pm
Location: Saigon, Vietnam

Re: Variable Percentage Withdrawal (VPW)

Postby AlohaJoe » Thu Mar 17, 2016 8:26 pm

longinvest wrote:In another thread, I stated that VPW eliminates sequence of return risk. It think that this bears repeating in this thread.

VPW eliminates sequence of return risk.


I don't understand how this is the case but maybe we have different definitions of sequence of returns risk.

I used the sequence of returns from Black Rock's paper on sequence of returns risk: https://www.blackrock.com/pt/literature ... -va-us.pdf where they show that Mrs. Doe, Mr. White, and Mr. Rush all have the same exact returns (so no market risk) but have different outcomes each year when withdrawing (due to sequence of returns).

I also added columns to show what the inflation-adjusted withdrawal is and then compare it to the $60,000 a year spending needs.

This is Mr. White's "bad sequence":

Image

This has a 91.5% failure rate. There are only 3 years that VPW provides for his spending needs. In some years he is up to $15,000 short and has to borrow money from his children just to pay his electricity bill.

Here is Mrs. Doe's "good sequence":

Image

This has an 8.5% failure rate. There are only 3 years that VPW fails to provide for her spending needs. And most years she has a sizeable surplus: she can go on an unexpected trip to Europe or give her children and grandchildren unexpected gifts.

Mr. White and Mrs. Doe had the same market risk, as Black Rock shows. I call the difference between Mr. White and Mrs. Doe "sequence of returns risk". They (and their extended family) live very different lives because of the sequence of returns, which VPW doesn't seem to protect them against.

One difference between VPW and constant withdrawal rates is that with constant ones the failures all happen at the end. Once you run out of money in year 32, then years 33, 34, 35, and so on will also (by definition) fail. With VPW, the failures can happen any year and still succeed the next.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Thu Mar 17, 2016 9:17 pm

AlohaJoe,

In the other thread, I have explained in details the difference between market risk and sequence of returns risk. Let me illustrate this with the example you have provided.

You have just shown a perfect example of how VPW does not exhibit any sequence of returns risk; how it only exhibit market risk.

Note how the withdrawals are identical at the end of each cycle of reordered returns for Mr. White and Mrs. Doe:

Code: Select all

Age White      Doe
65  $57,000.00 $57,000.00
70  $57,790.01 $57,790.01
75  $59,002.77 $59,002.77
80  $60,806.00 $60,806.00
85  $62,117.43 $62,117.43
90  $63,497.85 $63,497.85


You might also note that, at these ages, the portfolio balances are also identical.

If there was sequence of returns risk, these amounts would differ; but they don't because there isn't such risk with VPW.

At all other ages, both portfolios have experience different cumulative returns since retirement (market risk), and thus, withdrawals differ, as expected from a withdrawal method that adapts to market returns.

In other words: Different cumulative returns imply different VPW withdrawals. This has nothing to do with sequence!

Actually, you have just illustrated how VPW has completely eliminated sequence of returns risk, because it delivered the exact same withdrawals (and portfolio balances) every time both Mr. White and Mrs. Doe experienced the same cumulative return between retirement and the withdrawal (e.g. at ages 65, 70, 75, 80, 85, and 90).

Try the same example with the Constant-Dollar Withdrawal (SWR) method, and you'll see how they won't get the same portfolio balances (or withdrawals, in case of failure) at ages 70, 75, 80, 85, and 90 even though the cumulative returns are identical. Why different portfolio balances or withdrawals? Exactly because of sequence of returns risk.

Sequence of returns risk is about getting a different withdrawal (or portfolio balance) even though the cumulative return since retirement is identical, because the sequence of returns composing this identical cumulative return was different.

VPW is exposed to market risk. Market risk is about getting a lower withdrawal (and portfolio balance) when the cumulative return since retirement is lower, and a higher withdrawal (and portfolio balance) when cumulative return since retirement is higher.

But, all this is pretty theoretical. Market risk is important and can't be ignored.

To dampen market risk, it is suggested to apply VPW on a balanced portfolio (e.g. a portfolio with a significant amount of bonds) and to combine VPW withdrawals with life-long base income (Social Security possibly delayed until 70, pension if any, and if absolutely necessary an inflation-indexed SPIA).
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

BahamaMan
Posts: 896
Joined: Wed Oct 01, 2014 5:52 pm

Re: Variable Percentage Withdrawal (VPW)

Postby BahamaMan » Fri Mar 18, 2016 12:52 pm

AlohaJoe wrote:
I also added columns to show what the inflation-adjusted withdrawal is and then compare it to the $60,000 a year spending needs.

This is Mr. White's "bad sequence":

This has a 91.5% failure rate. There are only 3 years that VPW provides for his spending needs. In some years he is up to $15,000 short and has to borrow money from his children just to pay his electricity bill.



Let me highlight longinvest's explanation....The problem is not with the Withdrawal Method, but the decision to retire on $1Million and NEEDING $60 Grand a year. Pretty much a Bad plan all the way around.

There is such a thing as discretionary spending and non-discretionary spending. Make sure your Discretionary Spending is more than half of your withdrawal amount, so that you can cut your budget if your withdrawal amount is less. And Paying your electric bill is non-discretionary spending.

jjface
Posts: 2419
Joined: Thu Mar 19, 2015 6:18 pm

Re: Variable Percentage Withdrawal (VPW)

Postby jjface » Fri May 06, 2016 11:26 pm

I think it is more correct to say that the VPW method manages sequence of return risk rather than eliminate it. It does this by reducing withdrawal amounts. This is only going to work if you can always cover essential spending as BahamaMan pointed out. The example decided on an inappropriate minimum amount based on portfolio size. However it still remains that the withdrawals under VPW are lower under a bad sequence solely due to the sequence of returns as Aloha pointed out and not due to the size of the returns.

Actually, you have just illustrated how VPW has completely eliminated sequence of returns risk, because it delivered the exact same withdrawals (and portfolio balances) every time both Mr. White and Mrs. Doe experienced the same cumulative return between retirement and the withdrawal (e.g. at ages 65, 70, 75, 80, 85, and 90).


I am not sure I can draw that conclusion. In the example you simply have the same sequence but in 5 year intervals rather than 1 year intervals. This is why you have the same withdrawals every 5 years - there is no sequence of returns risk when the sequence is the same ie the 5 year sequence is identical ie 38% each 5 years for both Doe and White) Within the 5 year blocks you have different sequence of returns and hence are open to the sequence of return risk within these blocks which is being managed by reducing withdrawals by the VPW method.

Regardless this is all very useful and interesting - thank you for all your work on this method and anyone else who contributed.

longinvest
Posts: 2306
Joined: Sat Aug 11, 2012 8:44 am

Re: Variable Percentage Withdrawal (VPW)

Postby longinvest » Sat May 07, 2016 8:17 am

jjface wrote:I think it is more correct to say that the VPW method manages sequence of return risk rather than eliminate it. It does this by reducing withdrawal amounts. This is only going to work if you can always cover essential spending as BahamaMan pointed out. The example decided on an inappropriate minimum amount based on portfolio size. However it still remains that the withdrawals under VPW are lower under a bad sequence solely due to the sequence of returns as Aloha pointed out and not due to the size of the returns.

Actually, you have just illustrated how VPW has completely eliminated sequence of returns risk, because it delivered the exact same withdrawals (and portfolio balances) every time both Mr. White and Mrs. Doe experienced the same cumulative return between retirement and the withdrawal (e.g. at ages 65, 70, 75, 80, 85, and 90).


I am not sure I can draw that conclusion. In the example you simply have the same sequence but in 5 year intervals rather than 1 year intervals. This is why you have the same withdrawals every 5 years - there is no sequence of returns risk when the sequence is the same ie the 5 year sequence is identical ie 38% each 5 years for both Doe and White) Within the 5 year blocks you have different sequence of returns and hence are open to the sequence of return risk within these blocks which is being managed by reducing withdrawals by the VPW method.

Regardless this is all very useful and interesting - thank you for all your work on this method and anyone else who contributed.


Jjface,

I cannot change the fact that you personally define "sequence of returns risk" differently than how it is defined in the retirement literature. If you do not wish to use the traditional definition of "sequence of returns risk", then you will be able to claim all sorts of things about VPW and other withdrawal methods, but that will be misleading for others who assume that you are using the traditional definition.

Just consider this:
  1. Calibrate VPW for a 30-year withdrawal.
  2. Pick a big set of historical sequences of 29 inflation-adjusted portfolio returns.
  3. Simulate 30-year retirements using the chosen sequences of returns and VPW as withdrawal method. Note how after the 30th withdrawal, the portfolio balance is always exactly $0, and the portfolio is never prematurely depleted.
  4. Simulate 30-year retirements using the same sequences of returns and 4% SWR as withdrawal method. Note how after the 30th withdrawal, the portfolio balance is most often positive (and sometimes huge), but sometimes negative (indicating that the portfolio was prematurely depleted).

It is obvious that VPW displays no variability in final portfolio value while SWR displays wide variability in final portfolio value (and in successfully funding a 30-year retirement). We need to give a name to this variability, to this "risk".

In the 4% SWR retirement literature, using a simple mathematical construction, the name "sequence of return risk" was given to that risk. Here's the mathematical reasoning that lead to this name:
  1. Pick a set of 29 returns, such that a 30-year retirement using 4% SWR succeeds when these returns happen in a certain order and fails when they happen in a certain other order.
  2. Look at the outcome (success or failure) of all the possible sequences of these 29 returns.
  3. As the success or failure of a 30-year retirement changes based on the sequence of returns, without changing the set of returns, retirement researchers decided to call this risk "sequence of returns risk".

I know that you want to express that VPW is risky. I agree, it is risky. But, we have to give a non-ambiguous name to the risk contained in VPW. As I have just explained, it would be misleading to call this risk "sequence of returns risk". So, let's not call it that!

What should we call this risk? Here's a mathematical reasoning what will help us find a proper name for VPW's risk:
  1. Calibrate VPW for a 30-year withdrawal.
  2. Pick a specific withdrawal, N, within the 30 withdrawals, where N > 1 and N < 30. (N could be 5, for example, to indicate the 5th withdrawal).
  3. Pick a set of N-1 returns.
  4. Calculate the Nth withdrawal for all possible sequences of these returns. Note how the Nth withdrawal is always the same, regardless of the sequence.
  5. Note how the Nth withdrawal is determined by the cumulative time-weighted return of the portfolio since retirement.
  6. Note that Nth withdrawal might be sufficient or insufficient, to fully fund a retiree's expenses. The possibility of insufficiency is a risk.
  7. As this risk is related to the cumulative time-weighted return since retirement, call this risk "market risk".

Does this make any sense to you, now?
Bogleheads investment philosophy | Lifelong Portfolio: 25% each of (domestic/international)stocks/(nominal/inflation-indexed)bonds | VCN/VXC/VAB/ZRR

jjface
Posts: 2419
Joined: Thu Mar 19, 2015 6:18 pm

Re: Variable Percentage Withdrawal (VPW)

Postby jjface » Sat May 07, 2016 8:58 am

Okay sorry I see what you mean - I was never 100% sure of the definition but thought it would incorporate any risk related to the ordering of returns. Portfolio failure for many would mean a lower than required withdrawal as well which is probably where the difference of opinion comes in. But that is more to do with the inadequacy of the portfolio rather than the method.

Ideally it would be useful if there was a way to determine what the minimum level of withdrawal would be using VPW. Obviously it would have to be a guide using past data with the usual limitations. Then someone could say I am ready for retirement because my withdrawals using the method should never drop under what I ultimately need.

In the end everyone will face the same 'market risk'/'withdrawal size risk' (whatever you want to call it!) as most people would not be able to stomach a larger withdrawal when portfolio size is shrinking rapidly. In reality most would have to reduce withdrawals/expenses anyway unless they have some serious courage. I mean who would take $60k if a portfolio has shrunk to say $400k even if past history says it is okay?

Anyway keep up the good work!

skjoldur
Posts: 103
Joined: Thu Sep 25, 2014 3:11 pm

Re: Variable Percentage Withdrawal (VPW)

Postby skjoldur » Sat May 07, 2016 1:06 pm

I have a basic question about the recommended use of VPW which has to do with the standard advice to use it along with some kind of guaranteed floor.

First of all, I think we will be good candidates to use VPW for an earlyish retirement (~50). In particular, we are in a good position to live with highly variable expenses. This is due in part to temperament and in part to how much of our desired expenses are discretionary.

Our current expenses (adjusted for retirement with health ins. etc.) are about 2.5% of our current nest egg. In retirement I'd like to spend more, 3% would be pretty good, and 4% would be even more fun. On the other end, we could get by on ~1.7% with a bit of scrimping. VPW calculates an initial withdrawal of over 4% for year 1.

As an additional safety net, I am not counting SS or the fact that we could downsize from our HCOL location to free up more resources.

My question is: why not use VPW on the entire portfolio? Flooring seems very expensive and I'm not sure I even understand how we would add it (an inflation adjusted annuity at my age doesn't seem like a good idea).

Thanks for any comments!

dbr
Posts: 22163
Joined: Sun Mar 04, 2007 9:50 am

Re: Variable Percentage Withdrawal (VPW)

Postby dbr » Sat May 07, 2016 1:45 pm

skjoldur wrote:Thanks for any comments!


You addressed this issue in your own situation, but for the general situation a plan for retirement financing in which the income is variable out of control of the retiree has an obvious problem. One way to offset the problem of variable and unpredictable income is to dilute it with certain income from pensions, annuities, and Social Security. In the end there really is no way to completely avoid the problem of investment volatility if one chooses to use withdrawals from investments to fund retirement income.

skjoldur
Posts: 103
Joined: Thu Sep 25, 2014 3:11 pm

Re: Variable Percentage Withdrawal (VPW)

Postby skjoldur » Sat May 07, 2016 2:13 pm

Thanks dbr.

I guess I could make the question a bit more abstract and about VPW in general.

On the one hand, some happy implementers of VPW often suggest that in order to have sufficient funding for retirement (and for using VPW) that one should be able to handle large swings in spending to the tune of having discretionary spending be 1/2 your budget or more. This makes sense to me.

On the other hand, VPW is often recommended as just a portion of a complete retirement plan to be used in conjunction with a floor of some sort. I wonder though, if this approach of recommending a floor is inherently tied to VPW.

These two bits of advice seem additive, they both are quite risk reducing. Maybe one or the other is enough.

The issue of WR vs. floor and upside is a common theme on the forums. I'm sure I don't really understand the intricacies. But when I look closely I come away with the impression that a sufficiently large chuck of bonds in the AA is a sort of floor.

If you take the advice that VPW makes sense if you can handle large swings in your spending (say 50%) and if your AA is moderate (say 50/50) then it means that your bond allocation is roughly sufficient for 100% of your 'floor' expenses.

Maybe there is a rule of thumb in there somewhere?

BahamaMan
Posts: 896
Joined: Wed Oct 01, 2014 5:52 pm

Re: Variable Percentage Withdrawal (VPW)

Postby BahamaMan » Sat May 07, 2016 2:52 pm

skjoldur wrote:Thanks dbr.

I guess I could make the question a bit more abstract and about VPW in general.

On the one hand, some happy implementers of VPW often suggest that in order to have sufficient funding for retirement (and for using VPW) that one should be able to handle large swings in spending to the tune of having discretionary spending be 1/2 your budget or more. This makes sense to me.

On the other hand, VPW is often recommended as just a portion of a complete retirement plan to be used in conjunction with a floor of some sort. I wonder though, if this approach of recommending a floor is inherently tied to VPW.

These two bits of advice seem additive, they both are quite risk reducing. Maybe one or the other is enough.

The issue of WR vs. floor and upside is a common theme on the forums. I'm sure I don't really understand the intricacies. But when I look closely I come away with the impression that a sufficiently large chuck of bonds in the AA is a sort of floor.

If you take the advice that VPW makes sense if you can handle large swings in your spending (say 50%) and if your AA is moderate (say 50/50) then it means that your bond allocation is roughly sufficient for 100% of your 'floor' expenses.

Maybe there is a rule of thumb in there somewhere?



I am retired and using VPW..... I use my entire portfolio when calculating VPW. With that said my 'Floor' is my conservative Portfolio. I am 70% in Bonds. My Floor is also made up of Social Security, which I am delaying to age 70. My SS.

And I agree that your discretionary expenses should be less than half of your first VPW Withdrawal Amount. This will leave you with plenty of flexibility, if you need it. And think about it, if your Non-discretionary expenses equaled 4% SWR, it would not be a very fun retirement.

User avatar
BlueEars
Posts: 3471
Joined: Sat Mar 10, 2007 12:15 am
Location: West Coast

Re: Variable Percentage Withdrawal (VPW)

Postby BlueEars » Sat May 07, 2016 7:07 pm

Our portfolio is currently 60/40 and I like VPW as a basic guide. Would go to 40/60 when specific conditions are met. Would also go to all Treasury's for bonds under specific conditions. The floor I think of as being set by the final depletion age of something like 110. SS also helps for floor.

User avatar
gordoni2
Posts: 185
Joined: Wed Aug 15, 2007 12:20 am
Contact:

Re: Variable Percentage Withdrawal (VPW)

Postby gordoni2 » Fri May 20, 2016 4:54 pm

Here are some thoughts on variable withdrawals without annuities or a bequest motive. The following strategies are all derived by defining the problem mathematically and then computing the mathematically optimal withdrawal strategy for that problem.

Fixed lifespan, no returns - 1/N. This is the simplest approach.

Fixed lifespan, fixed returns - VPW. The problem can be solved by working backwards, or in closed form as consumption = r . (1+r)^(N-1) / ((1+r)^N - 1) where r is the real rate of return and N is the number of years remaining.

Fixed lifespan, fixed returns, risk aversion - "FLRA". A hidden assumption of VPW is that the risk aversion is infinite. Risk aversion is how important it is have smooth consumption. Reducing risk aversion means you are willing to experience slightly lower consumption now in order to allow the portfolio to grow and experience higher consumption in the future. In the case of constant relative risk aversion this problem can be solved as a special case of Merton's portfolio problem, or as a special case of the next problem.

Stochastic lifespan, fixed returns, risk aversion - "SLRA". A fixed lifespan isn't realistic. It results in assigning an equal weight to well being at age 50 and 99 even though you are unlikely to make it age 99. I haven't seen this problem explored previously, but for constant relative risk aversion I think I was able to solve it by searching for the optimal consumption level at each age and working backwards.

Fixed lifespan, stochastic returns, risk aversion - Merton's portfolio problem (MPP). This requires a risk free asset class, constant relative risk aversion, and the adoption of the optimal asset allocation, which interestingly is fixed for all ages and portfolio sizes. It is generalized to multiple asset classes in Merton's original paper.

Stochastic lifespan, stochastic returns, risk aversion, guaranteed income - stochastic dynamic programming (SDP). A hidden assumption of all the finite risk aversion problems is that there is no guaranteed income, which isn't realistic. This is the most complex of the problems. It can be generalized to multiple asset classes. It can be solved numerically by working backwards. The downside is the solution is complex to develop and explain, it might take 5-10 seconds to solve even on a fast desktop computer today, and it is very situation specific and so difficult to present in tabular form.

Now the hard part. Which strategy to use? It would seem the more realism the better, and SDP is the most realistic. Unfortunately the complexity of SDP is often a problem. Choosing between the remaining schemes is difficult. Part of the beauty of all variable withdrawal schemes though is they are adaptive so a slightly sub-optimal decision at one age doesn't greatly mess up decisions at subsequent ages. I haven't done so, but it would be instructive to compute certainty equivalent consumption levels associated with each of these schemes for representative scenarios.


Return to “Investing - Theory, News & General”

Who is online

Users browsing this forum: Bing [Bot] and 24 guests