Is anyone using Boglehead's Variable Withdrawal Percentage?

[Koolau]

Our withdrawal varies each year and never bother with withdrawal percentage.

Yeah, I calculate in hindsight just to be certain I'm not overspending. Even if I am, I can correct the next year. That actually only happened the first few years when I was getting established in a new (HCOL) area.

 

[RetiredHappy]

Yeah, I calculate in hindsight just to be certain I'm not overspending. Even if I am, I can correct the next year. That actually only happened the first few years when I was getting established in a new (HCOL) area.

As long as my investments are growing above inflation rate over several years, I don't care how much I am withdrawing.

 

[Sunset]

I know we aren’t supposed to mention another tool, but if you use FICalc, not FireCalc, it gives you the ability to explore several withdrawal strategies including the Vanguard variable method. Here is a list of all the strategies that you can explore in the tool. It’s free.

Thanks.... it's interesting to play with.

 

[Dtail]

We use the TPAW Planner developed by Ben Matthews. There's a very long and excellent ongoing thread about it on Bogleheads but here's the link to it. The introduction and tutorials should help you decide if it's for you.

https://tpawplanner.com/

IMHO both VPW and TPAW are far superior to SWR approaches of any sort because there's zero chance of running out of money. The hardest part of making the change is learning to think in terms of amortization rather than just winging it.

Firecalc also has a variable WR% module where one would never run out of money.

 

[lem1955]

I used VPW from age 60 to 68 and spent an average 5.5% of portfolio for those years. We had lived in our house 25-33 years and it needed a kitchen upgrade, new windows and a new roof. And we needed new cars in that period. I was holding off on taking SS until 70 and expected some inheritance (at least $500k). Actually received closer to $800k in 2023. This year I am 70 and just received first SS deposit. WR is now below 2%.

 

[big-papa]

Firecalc also has a variable WR% module where one would never run out of money.

Where is that? The only thing that I see that would never run out of money is the fixed percentage of portfolio method on Firecalc. Mathematically, that will not run out of money but you have to choose the percentage somehow and hope it's not too high. While, mathematically, you can't run out of money, you can certainly asymptotically approach it.

Cheers.

 

[deserat]

Excellent discussion. Interestingly, the originator of TPAW had a long online discussion with the originator of VPW. As with any tool, one needs to know the best scenario for that tool and its limitations. I spent a *lot* of time looking at retirement spending models before I retired. I decided to use VPW when I do withdraw from my portfolio merely because it provided some guardrails, allowed for easy input of my situation with regard to pensions and investments and is built with enabling adaptation to market conditions while advocating for a 'floor' for lifestyle costs by purchasing an annuity at 80 (which is not necessary if one has a pension which can cover that). I also like that the tool is being forward tested with actual market conditions. I'm also more of a 'hatchet' versus scalpel person with regard to managing my money, ie, I've probably over-saved for my lifestyle requirements, so can afford some 'slop' in the measurements. VPW allows me to have some rigor, gives me a range to withdraw and lets me get on with living my life. As one commenter stated, one can use a smoothing process in conjunction. KlangFool, who is quite risk averse, has a much more conservative smoothing mechanism that is based on two years and 'safe' investments while using VPW. One can slice and dice how they like. BL: I appreciate the work in developing the model and spreadsheet and find it is a tool that meets my needs.

 

[Dtail]

Where is that? The only thing that I see that would never run out of money is the fixed percentage of portfolio method on Firecalc. Mathematically, that will not run out of money but you have to choose the percentage somehow and hope it's not too high. While, mathematically, you can't run out of money, you can certainly asymptotically approach it.

Cheers.

Percentage of remaining portfolio is the module. If one sets it to 95%, it is effectively the Clyatt 95/5 variable WR%. This is a variable variation on the 4% guidance.
Yes conceptually one can approach zero I suppose like one can approach zero in the VPW methodology.

 

[big-papa]

Percentage of remaining portfolio is the module. If one sets it to 95%, it is effectively the Clyatt 95/5 variable WR%. This is a variable variation on the 4% guidance.
Yes conceptually one can approach zero I suppose like one can approach zero in the VPW methodology.

OK, that's what I thought. It's a variable dollar methodology not a variable percentage methodology. 100% agree that this can happen with VPW. And one of the reasons that it can happen with VPW is that the author of it choose fixed rates for the amortization for both stocks and for bonds. As I noted upstream, there are a number of different valid ways to estimate stock returns, though of course none of them will be great. A fixed rate based on a worldwide, long term trend is one, such as what VPW uses. If you've ever seen such a plot, there does seem to be something of a straight line (on a log plot) return trend. But over short or even intermediate timeframes, stock returns can drop below that trendline. When that happens, VPW over-withdraws. This reduces the portfolio value even more. And that results in smaller withdrawals. The cycle continues until returns finally make their way back above the trendline. This is what would have happened to a late 60's retiree who used this method if they had a stock heavy portfolio - that is a "U" shaped withdrawal curve. Using valuations to update expected returns would have done slightly better but only slightly - that's because, at best, their accuracy for returns is best over, say, 10 years or so. Not great but better than a fixed trend line, on average.

Bonds, on the other hand are different. Again, if you look at bond fund yields over the longer term, there is no long term straight trendline like stock has. On the other hand, if you know the yield today and have a decent inflation expectation measurement, you can get a pretty good prediction of real yield between now and the duration of the bond fund. Or use a TIPS fund directly and use its real yield. Updating the expected returns before each withdrawal calculation will give you smoother withdrawals for the bond portion of a portfolio. Backtesting this back during the 1970's is difficult because TIPS didn't exist and their one of the tools that are used to create inflation expectations.

An amortization system that regularly updates expected returns of stocks and bonds (or at least bonds), is much less likely to result in a withdrawal that asymptotically can approach zero, while giving you the higher average withdrawals in real dollars over the designated length of retirement.

Cheers.

 

[big-papa]

Excellent discussion. Interestingly, the originator of TPAW had a long online discussion with the originator of VPW. As with any tool, one needs to know the best scenario for that tool and its limitations. I spent a *lot* of time looking at retirement spending models before I retired. I decided to use VPW when I do withdraw from my portfolio merely because it provided some guardrails, allowed for easy input of my situation with regard to pensions and investments and is built with enabling adaptation to market conditions while advocating for a 'floor' for lifestyle costs by purchasing an annuity at 80 (which is not necessary if one has a pension which can cover that). I also like that the tool is being forward tested with actual market conditions. I'm also more of a 'hatchet' versus scalpel person with regard to managing my money, ie, I've probably over-saved for my lifestyle requirements, so can afford some 'slop' in the measurements. VPW allows me to have some rigor, gives me a range to withdraw and lets me get on with living my life. As one commenter stated, one can use a smoothing process in conjunction. KlangFool, who is quite risk averse, has a much more conservative smoothing mechanism that is based on two years and 'safe' investments while using VPW. One can slice and dice how they like. BL: I appreciate the work in developing the model and spreadsheet and find it is a tool that meets my needs.

Yep, I remember that thread very well. Challenging anything about VPW usually gets you a response (sometimes pre-canned) from the author, and I think that's one of the things that started the entire conversation the two of them had.

I probably spent an entire decade looking at various withdrawal methods before I retired. Like most, I started with SWR (and many of what I call "hacks" to SWR which attempt to improve it). Some of them can be algorithmically intense.

Ultimately, I ended up with amortization. I really like TPAW but by the time it was in a decent state I was too far along with my plan to incorporate it. Instead, I'm living vicariously through my daughter who is following something very similar to it after she graduated university a couple of years ago.

I started retirement with amortization using duration matched/targeted TIPS funds and stocks funds and withdrawing from both using amortization, making quarterly withdrawals. For TIPS I calculated the aggregate real yield and updated the average duration of the two to match my investment horizon each quarter. For stocks I used 1/CAPE as a crude estimate for future real returns. I stopped rebalancing between stocks and bonds. I also used Time Value of Money concepts to withdraw "extra" from my bond funds for future SS streams by calculating their net present value and adding that to the value of my bond funds. Putting together the spreadsheet was not difficult, nor was deriving a few of the formulas. Once that was figured out, the process each quarter was looking up the data and just plugging it into my spreadsheet. That took all of about 10 minutes and then I made the withdrawals and got on with things.

The net result was that we were withdrawing a lot more than we were able to spend. This year I took our TIPS bond funds, sold them, and purchased ladders. coupons + maturing bonds in the ladder along with dividends thrown off by our stock funds now easily pay for all of nondiscretionary spending, without a lot of excess. I now use amortization to calculate a max withdrawal we could make by selling shares of our stock funds each year. The only regular withdrawals we're making from that so far, is extra to pay tax on the Roth conversions we make each December and for the next few years until that's done. That, by itself, doesn't put much of a dent in the value of our stock funds, fortunately.

For a very good 2 part series on amortization, including time value of money concepts, along with a sample spreadsheet, I highly recommend this two part series on the bogleheads blog by user siamond.
(If you're not into spreadsheets, feel free to ignore all of it)

Early Retirement and Time Value of Money (Part 1) – Financial Page

Early Retirement and Time Value of Money (Part 2) – Financial Page

Cheers.

 

[audreyh1]

Where is that? The only thing that I see that would never run out of money is the fixed percentage of portfolio method on Firecalc. Mathematically, that will not run out of money but you have to choose the percentage somehow and hope it's not too high. While, mathematically, you can't run out of money, you can certainly asymptotically approach it.

Cheers.

With FIREcalc can look at various allocations and see how far you might draw down. Also which ones seems sustainable over long periods. For 50/50 AA it was something like 4.35% with worse case drawdown of 55% inflation adjusted, and average inflation adjusted ending portfolio equal to starting portfolio.

I don’t draw that high a percentage yet.

 

[big-papa]

With FIREcalc can look at various allocations and see how far you might draw down. Also which ones seems sustainable over long periods. For 50/50 AA it was something like 4.35% with worse case drawdown of 55% inflation adjusted, and average inflation adjusted ending portfolio equal to starting portfolio.

I don’t draw that high a percentage yet.

Yes I looked at this method in the past, extensively. Ultimately I came to the conclusion that future sequences of return have no obligation to be encompassed by past sequences of returns. As a result I put this to the side.

There is always going to be some sort of looking at the past in any withdrawal method. My goal was to find something that abstracted the past as much as possible. That's not to say that there is anything that's risk-free. That doesn't exist.

Cheers

 

[pjigar]

Each of my kids gets up to $38K each year. We use $38K because that is what my wife and I can give each child each year ($19K*2). Our thoughts on this developed out of reading Bill Perkin's "Die with Zero, how to get all you can from your money and your life".

Anyone who expects to leave an inheritance should take advantage of the annual tax-free gift limit: Tax-free to you and the recipient, reduced eventual estate and potential estate tax, and it would be vastly more appreciated by the recipient compared a larger inheritance later.

We layer one more benefit on top of the ones I mentioned earlier: wealth building journey for the recipient. We don't just "give" annual gift but rather "match" 100% of recipient's retirement contributions (up to annual tax-free gift amount). We call it "Bank of Mom and Dad matching contributions"! Recipient becomes a regular investor AND they get "free" money as a gift. They can have their cake and eat it too.

 

[RunningBum]

Anyone who expects to leave an inheritance should take advantage of the annual tax-free gift limit: Tax-free to you and the recipient, reduced eventual estate and potential estate tax, and it would be vastly more appreciated by the recipient compared a larger inheritance later.

A disadvantage to this is that gifted assets keep the original cost basis you had, while inherited assets get stepped up basis. That said, I am gifting my son the tax-free gift limit annually. As you say, it's much more useful to him now. At some point I may stop because of the basis reason, or if it turns out I need the money more than he does.

 

[audreyh1]

Yes I looked at this method in the past, extensively. Ultimately I came to the conclusion that future sequences of return have no obligation to be encompassed by past sequences of returns. As a result I put this to the side.

There is always going to be some sort of looking at the past in any withdrawal method. My goal was to find something that abstracted the past as much as possible. That's not to say that there is anything that's risk-free. That doesn't exist.

Cheers

Since the data went back to 1871 and those first 60 years covered some massive boom and bust cycles, I decided that was good enough for me. Of the 12 worst case scenarios I reviewed for the %remaining portfolio withdrawal method, 11 of them occurred during that period. Ignoring that early period meant worst case drawdown was only 50% instead of 55%. So IMO including that gave me an extra 5% worst case to cover much worse future scenarios. If it’s something more catastrophic than that, well IMO you can’t really plan for it, you have to take each year as it comes and do your best, everyone will be seriously hurting.

Practically speaking however, that was just an eagle eye view type exercise for a particular allocation. I know how robust the method is, I have a good idea how things could unfold - generally very gradually. And how lower than max (max = the sustainable rate) withdrawal rates play out on average.

 

[audreyh1]

A disadvantage to this is that gifted assets keep the original cost basis you had, while inherited assets get stepped up basis. That said, I am gifting my son the tax-free gift limit annually. As you say, it's much more useful to him now. At some point I may stop because of the basis reason, or if it turns out I need the money more than he does.

We gift cash anyway, which is generally from funds we didn’t otherwise spend, so we don’t worry about gifting securities or stepped up basis.

 

[pjigar]

A disadvantage to this is that gifted assets keep the original cost basis you had, while inherited assets get stepped up basis. That said, I am gifting my son the tax-free gift limit annually. As you say, it's much more useful to him now. At some point I may stop because of the basis reason, or if it turns out I need the money more than he does.

Yes, you are correct about the basis. But we expect to blow past the estate tax exemption limit so every annual tax-free gift is saving away 40% eventual estate tax. YMMV.

 

[TheWizard]

We gift cash anyway, which is generally from funds we didn’t otherwise spend, so don’t worry about gifting securities or stepped up basis.

Same here. Recurring gifts to my offspring come from excess Ordinary Income. Remaining excess Ordinary Income gets invested in my taxable account which may have stepped up basis down the road, unless I spend it first.

Charitable gifts, OTOH, are made from my tIRA with QCDs...

 

[Dtail]

Since the data went back to 1871 and those first 60 years covered some massive boom and bust cycles, I decided that was good enough for me. Of the 12 worst case scenarios I reviewed for the %remaining portfolio withdrawal method, 11 of them occurred during that period. Ignoring that early period meant worst case drawdown was only 50% instead of 55%. So IMO including that gave me an extra 5% worst case to cover much worse future scenarios. If it’s something more catastrophic than that, well IMO you can’t really plan for it, you have to take each year as it comes and do your best, everyone will be seriously hurting.

Practically speaking however, that was just an eagle eye view type exercise for a particular allocation. I know how robust the method is, I have a good idea how things could unfold - generally very gradually. And how lower than max (max = the sustainable rate) withdrawal rates play out on average.

Agree. Additionally IMHO I am not sold on the VPW method being any better than the Clyatt 95/5 or a simplified % of remaining portfolio.
None of the worst historical starting retirement scenarios have happened in over 55 years.

 

[audreyh1]

Agree. Additionally IMHO I am not sold on the VPW method being any better than the Clyatt 95/5 or a simplified % of remaining portfolio.
None of the worst historical starting retirement scenarios have happened in over 55 years.

A benefit of VPW is that it reduces the years you have left, so you can spend down faster as you get older. It’s going to be more “efficient” in terms of spending while you are alive. However, folks who use VPW generally set a high age of death like 110, because they want some cushion.

My approach to shrinking life span as we age is to rarely increase the % we withdraw. For example, we were using 3% for a very long time, and since DH turned 70 this year, I decided to bump it up to 3.5%, still relatively conservative. Another 5 - 10 years we’ll reevaluate. We also don’t mind having a big buffer to fund higher late in life expenses, care, etc.

 

[Dtail]

A benefit of VPW is that it reduces the years you have left, so you can spend down faster as you get older. It’s going to be more “efficient” in terms of spending while you are alive. However, folks who use VPW generally set a high age of death like 110, because they want some cushion.

My approach to shrinking life span as we age is to rarely increase the % we withdraw. For example, we were using 3% for a very long time, and since DH turned 70 this year, I decided to bump it up to 3.5%, still relatively conservative. Another 5 - 10 years we’ll reevaluate. We also don’t mind having a big buffer to fund higher late in life expenses, care, etc.

So if one is setting the death cushion to 110, then it mathematically/theoretically cancels out any advantage to remaining years of life.
Agree with the increase in % withdrawal as an alternate way of maximizing usage of portfolio as one gets older.

 

[audreyh1]

So if one is setting the death cushion to 110, then it mathematically/theoretically cancels out any advantage to remaining years of life.
Agree with the increase in % withdrawal as an alternate way of maximizing usage of portfolio as one gets older.

Well it’s really about how much cushion you want anyway, and/or how much you would like to pass on to heirs.

The reevaluation to increase withdrawal % every 5 years or so is nice and simple as is the otherwise fixed % every year.

 

[big-papa]

Since the data went back to 1871 and those first 60 years covered some massive boom and bust cycles, I decided that was good enough for me. Of the 12 worst case scenarios I reviewed for the %remaining portfolio withdrawal method, 11 of them occurred during that period. Ignoring that early period meant worst case drawdown was only 50% instead of 55%. So IMO including that gave me an extra 5% worst case to cover much worse future scenarios. If it’s something more catastrophic than that, well IMO you can’t really plan for it, you have to take each year as it comes and do your best, everyone will be seriously hurting.

Practically speaking however, that was just an eagle eye view type exercise for a particular allocation. I know how robust the method is, I have a good idea how things could unfold - generally very gradually. And how lower than max (max = the sustainable rate) withdrawal rates play out on average.

The thing with all of these methods that depend on past sequences of returns to determine suitability is that I am able to create worse sequences that are worse and not even apocalyptic in nature.

Cheers

 

[RobLJ]

Along the lines of audrey1, I planned taking out above 4% (two years were close to 6% withdrawal) until SS, when the rate would go down to a more "reasonable" rate of 4 or under. So, this year I'm taking SS and the withdrawal rate originally was 3.5%, but I upped it to 4.5%.
To be honest, we haven't been spending everything we withdraw, so we're probably going to use some of the overage stuffed in the brokerage fund in SNSXX to buy a used camper van next week (55k). For now, I'll keep withdrawing from the IRAs and 403bs up to 4.5-5% and just keep what is unspent in the brokerage. When DW begins taking her SS, then we'll really have to up our spending game, assuming the market holds up and the crick don't rise.

 

[big-papa]

A benefit of VPW is that it reduces the years you have left, so you can spend down faster as you get older. It’s going to be more “efficient” in terms of spending while you are alive. However, folks who use VPW generally set a high age of death like 110, because they want some cushion.

My approach to shrinking life span as we age is to rarely increase the % we withdraw. For example, we were using 3% for a very long time, and since DH turned 70 this year, I decided to bump it up to 3.5%, still relatively conservative. Another 5 - 10 years we’ll reevaluate. We also don’t mind having a big buffer to fund higher late in life expenses, care, etc.

I think that the vpw spreadsheet does preprogram what I believe is an excessively high max age in it but that can be changed.

With the amortization method I use, I stop at age 95. I think it's reasonable but one should still have a plan B. We have a couple. First is that amortization doesn't require you have a $0 terminal portfolio value. You can make it anything you want but you should probably increase it by the rate of inflation. That can be tapped by reducing the terminal value and extending the timeframe for withdrawals. Secondly we have I-Bonds maturing every year starting in our mid 80's for a total of 10 years which can be used to extend our TIPS ladder, purchase a SPIA or whatever.

Finally a lot of us use the calculated withdrawal only for an upper bound. You don't have to withdraw the full amount if you don't need it all. Anything not withdrawn just increases all future withdrawal capacity.

Cheers