Revisiting the asset allocation question - The case for 100% stocks

[Interest Compound]

I’ve always been a staunch advocate for bonds in any portfolio. But I’m on the verge of switching to 100% stocks, and want to share my reasoning. I'm sure a lot of this is already obvious to many of you, so feedback is welcome! :)

VPW

Variable Percentage Withdrawal (VPW) initially drove me to consider 100% stocks, as that allocation supports a higher withdrawal rate. It can do this, because unlike the 4% rule, it aims to draw down your portfolio. A 30 year old using VPW can initially withdraw 5.1% with 100% stocks, but change it to 80/20 stocks/bonds and the initial withdrawal drops to 4.5%, meaning I'll have to work longer to reach FIRE.

Discovering VPW removed a big hurdle of 100% stocks for me. The fear of portfolio failure is gone, as VPW cannot fail. It tells you exactly how much you can safely withdrawal, there is no risk of prematurely depleting your portfolio, you end up with much more money to spend than simply sticking to the 4% rule, I don’t have to live paycheck to paycheck in retirement, and I can FIRE earlier. I discuss the pros and cons here:

http://forum.mrmoneymustache.com/post-fire/playbook-on-down-marketsportfolio-steps/

Here's a breakdown of the difference between VPW and the 4% rule for a 30 year old early retiree, with a starting retirement year of 1929:

VPW is the solid line, 4% rule is the dotted line.



We see the 4% rule portfolio was wiped out after 22 years of retirement, while VPW grew to 3.5 million, before completing the drawdown on schedule after 70 years of withdrawals. The last 20 years of withdrawals averaged $400,000 a year.

Starting year of 1966:



We see the 4% rule portfolio was wiped out after 25 years of retirement, while VPW grew to 7 million, and still has 20 years of withdrawals to go.The current-year withdrawal is $400,000.

Long story short, I’m ok with the income fluctuations inherent in VPW. None of the worst case historical scenarios look scary to me, and stocks did better than bonds during those periods with VPW anyway. I think it's important to be flexible and I'm careful not to confuse retirement expense needs with withdrawal amount. Expense needs are part of budgeting. withdrawal amount is like an income. When you were working you had a budget and an income. I don't think retirement should be any different. I didn't live paycheck to paycheck during my working years, and I don't want to live paycheck to paycheck in retirement. We always warn people of the risks when living paycheck to paycheck during their working years, but if your expenses are exactly 40k, and you save up exactly 1 million and retire with the 4% rule, that looks very much like living paycheck to paycheck to me. That's more risk than I'm willing to take. The easiest way for me to avoid this, is to increase my income with side-gigs…by doing all the fun stuff I'll be doing anyway.

I genuinely can't imagine not doing anything that generates income for 50 years. Not because I'm a workaholic, but because it seems to be the natural progression of learning new skills. Suddenly you have tons of free time to FINALLY work on the things your passionate about. Maybe you want to write, learn a programming language, design websites, learn photography, build houses (MMM)...etc. Chances are, someone out there is willing to throw money at you to develop those new skills.

I have tons of friends who decided to learn photography. Within a month they were booking paid gigs. Two paid gigs a month, as a beginner doing something they love, is enough to bring you from a 4% withdrawal rate, to 3% on a $1,000,000 portfolio.

That said, with VPW, chances are high the side-gigs won’t be necessary. VPW gives me a higher paycheck in the vast majority of years, and a lower paycheck in the rare times when I'm in danger of a portfolio wipe. The 4% rule is like VPW, but with a fixed paycheck. Sure it's easier to budget, but it's the worst of both worlds. You miss out on the higher paycheck in the vast majority of years, and you still risk a portfolio wipe. It's easy to look at VPW and be afraid of "a lot of variability". But in your working career, if you were offered a 50% raise would you turn it down because it had "a lot of variability"? I like a lot of variability on the upside, I have planned for the variability on the downside.

Stock ladder?

I read an article, can’t find it now, that presented investing in a way I hadn’t thought of it before then. Imagine someone saving up for retirement, between the ages of 30 through 60. They said we should visualize it like a CD ladder:

The deposits of age 30 ———> will fund the expenses of age 60
The deposits of age 31 ———> will fund the expenses of age 61
The deposits of age 32 ———> will fund the expenses of age 62
…
The deposits of age 58 ———> will fund the expenses of age 88
The deposits of age 59 ———> will fund the expenses of age 89
The deposits of age 60 ———> will fund the expenses of age 90

Looking at it this way, each deposit has a full 30 years to “cook”.  A full 30 year time horizon, before it gets used up. As we all know, volatility in stock returns over 30 years gets very narrow…similar to bonds actually…just a bit further up the Y axis:



Bonds:



Even if stocks drop 50% the year you turn 60, only the deposit you made when you were 30 years old is affected, because that’s the only deposit you’re withdrawing that year. Given this visualization, it seems like a no-brainer to go 100% stocks. Of course I'll be FIRE much earlier than 60 years old, but the vast majority of my money will have decades to "cook" before being used.

While doing research on the topic, Brooklynguy had some really insightful posts along these same lines:

--------------------------------------------------------------------
But the "specific years" you need stocks to outperform bonds are not the years between investment and the commencement of retirement, but the entire remainder of your life.  Even during the accumulation phase when you're working towards the goal of accumulating enough to pull the trigger on retirement, the time horizon for your investments is the rest of your life (not your retirement date).
--------------------------------------------------------------------
Source

--------------------------------------------------------------------
Of course 100% stocks doesn't make sense if your investing time horizon is less than a decade!  But in reality, when you invest during the accumulation phase, your investment time horizon is the entirety of the remainder of your life (i.e., multiple decades) (because you are not going to be accessing those funds immediately upon retirement--you are going to be slowly drawing them down over the rest of your life).

The historical data show that the optimal asset allocation for multi-decade periods was always 100% stocks (or, depending on the precise settings for variables such as your investment expense ratio, near 100% stocks)--this allocation produced both the highest portfolio values and the highest success rates in every historical period covered by cfiresim.  So, if you assume that the future will be no worse than the past, then 100% stocks is (beyond argument) the best allocation.
--------------------------------------------------------------------
Source

Emergency fund - break even calculation

After seeing the break-even calculation on investing vs keeping an emergency fund in cash, I’ve never held an emergency fund. In short, money in stocks is expected to just about double every 8 years or so. After that point, even if stocks crash 50%, you'll still be ahead by investing it vs keeping it in cash. In that sense, a cash emergency fund really has a limited shelf-life of usefulness:

(Green means you were better off investing the money, even after the crash)



For the above calculation, I used 10.2% for stocks, and 1% for the savings account:



While these numbers change wildly in reality, the general premise should hold true.

Bonds - break even calculation

I never thought to apply the same calculation to bonds! It seems bonds and emergency funds, both have a limited shelf-life:



I used 5.5% for bonds:



Withdrawals change the chart

Withdrawals change the above charts pretty significantly. Here's a 100% stock vs 100% bond chart with no withdrawals. We see they touch during the 2008 crash:



Now if we start with a $1,000,000 portfolio, and $40,000 yearly deposits...



Since stocks spend most of their time growing, then crash quickly, most of your withdrawals will be taken during up years. As a result, the withdrawals don't depress the stock portfolio as much, creating a bigger buffer before the 2008 crash.

GoCurryCracker's path to 100% equities

This article was the final push I needed. "Wait a second?!  Isn’t 100% stock super risky?  What if the stock market suffers a major drop?  What if the Great Recession happens all over again? I think the phrasing of these questions holds an underlying assumption, that the stock price on any given day is important.  Unless you plan to buy or sell, the price is largely irrelevant."

Staying the course

I've been investing for 15 years, and am quite comfortable with staying the course during a crash. Please consider the following scary links before making a similar move:

What was the 2008 crash like in real time?
http://www.livingafi.com/2014/05/drawdown-part-1-the-basics/
http://forum.mrmoneymustache.com/investor-alley/why-would-i-be-in-anything-other-than-100-stocks/msg541017/#msg541017
http://forum.mrmoneymustache.com/investor-alley/why-would-i-be-in-anything-other-than-100-stocks/msg541078/#msg541078

Thoughts?

[arebelspy]

It's obvious you put a lot of work into this post, so thanks.

I remain skeptical of basing any retirement decisions on the VPW, as per our discussion in the other thread on it.

I go back and forth on 100% stocks (or maybe 90/10 at the most) versus a much more diversified (including gold, etc., more PP based).  I personally wouldn't go 60/40 or anything like that--high stocks, or high diversification into various assets, but just a middling stocks/bonds seems suboptimal to both of those options to me.

[BlueMR2]

Mirrors a lot of research I've been doing recently that makes bonds look like total losers (other than as a security blanket), but only in the case where one does not rebalance.  There *appear* to be real gains to be had long term with at least a little bit of bonds available to rebalance into (sort of an automatic way to do some actual, working market timing).  I didn't see that in your calculations, did I miss it, or did you find it to not be worthwhile?

[Ktfeehan]

I absolutely agree with 100% stock approach during retirement assuming a paid off home and a cash emergency fund.  Together they are major stabizing factor.  I almost never see this asked or taken into account when discussing asset allocation % mixes.

[brooklynguy]

The historical data show that the optimal asset allocation for multi-decade periods was always 100% stocks (or, depending on the precise settings for variables such as your investment expense ratio, near 100% stocks)--this allocation produced both the highest portfolio values and the highest success rates in every historical period covered by cfiresim.  So, if you assume that the future will be no worse than the past, then 100% stocks is (beyond argument) the best allocation.

Because I can already see Tyler (creator of PortfolioCharts.com) wincing at these statements, let me clarify that they, like the original post in this thread, were referring to asset allocation as a sliding scale solely between two options:  "stocks" and "bonds," each as a single unified asset class.  In reality, per Rebs' comment above regarding more diversified portfolios (like the permanent portfolio), there are of course specific asset allocations that have fared better historically once you start slicing and dicing these two generic asset classes into more specific subclasses and/or add other asset classes into the mix.  (My own favored approach continues to be to keep things simple and stick with 100% stocks, using a portfolio comprising broadly-diversified domestic and international total market index funds.)

[Interest Compound]

Mirrors a lot of research I've been doing recently that makes bonds look like total losers (other than as a security blanket), but only in the case where one does not rebalance.  There *appear* to be real gains to be had long term with at least a little bit of bonds available to rebalance into (sort of an automatic way to do some actual, working market timing).  I didn't see that in your calculations, did I miss it, or did you find it to not be worthwhile?

Everything I've found shows a reduction in returns, and therefore withdrawals, with any allocation to bonds. Here's cfiresim (Custom VPW) with 100% stocks:



and 90% stocks:

[marty998]

Good post, nice to read a new/different viewpoint on an old topic.

[rocket354]

Hear, hear. All of these discussions of course rely on the assumption that returns going forward will be similar to historical returns. But with that assumption, the numbers don't lie. I'm at 100% stocks right now and don't see changing that.

The rationalizations for people getting into lower-return investments (or investing sub-optimally, eg, DCA) need to just go away. The real reason is people are scared. Anything else is just an attempt to justify that fear.

[Tyler]

The historical data show that the optimal asset allocation for multi-decade periods was always 100% stocks (or, depending on the precise settings for variables such as your investment expense ratio, near 100% stocks)--this allocation produced both the highest portfolio values and the highest success rates in every historical period covered by cfiresim.  So, if you assume that the future will be no worse than the past, then 100% stocks is (beyond argument) the best allocation.

Because I can already see Tyler (creator of PortfolioCharts.com) wincing at these statements, let me clarify that they, like the original post in this thread, were referring to asset allocation as a sliding scale solely between two options:  "stocks" and "bonds," each as a single unified asset class.  In reality, per Rebs' comment above regarding more diversified portfolios (like the permanent portfolio), there are of course specific asset allocations that have fared better historically once you start slicing and dicing these two generic asset classes into more specific subclasses and/or add other asset classes into the mix.  (My own favored approach continues to be to keep things simple and stick with 100% stocks, using a portfolio comprising broadly-diversified domestic and international total market index funds.)

Good to see you know me well.  ;)  Your comments are spot-on.  (And your own plan seems perfectly reasonable.)

I appreciate the well-researched post, Interest Compound.  Just a few comments:

1) Yes, when you limit your investing options to a single broad stock fund and a broad bond fund, high percentages of stocks usually win out.  Interestingly, research has shown that 100% stocks has a lower SWR than lower stock allocations, and there's no real difference in retirement safety between 35% stocks and 80% stocks.  The sweet spot for maximizing both withdrawal rates and long-term returns for a traditional stock/bond portfolio seems to be about 75-80% stocks.  Importantly, other portfolios that use more diverse funds have different results.  (more on that below)

From Wade Pfau


2) Specifically referencing the charts you show from PortfolioCharts, the following comment jumped out: "As we all know, volatility in stock returns over 30 years gets very narrow…similar to bonds actually…just a bit further up the Y axis".  I think I know what you're getting at, but to be clear for those interpreting the charts: The uncertainty of the long-term average return gets lower the longer you hold an investment, but the annual volatility is exactly the same in year 30 as it is in year 1.  The distinction is very important for something like withdrawal rates.   Despite the 3% difference in CAGR (since 1972) for the two portfolios you show, the difference in long-term SWR over that same timeframe is less than 1%.  The high volatility of stocks negates much of the average returns benefit when considering drawdowns, and it's absolutely possible for a lower-return portfolio to be much safer in retirement than a higher-return option depending on the volatility.  Here are a few examples.  Note how the 60-40 and TSM (total stock market) pretty much match the chart above, and also how none of the top 3 SWR portfolios have more than 40% stocks.

From PortfolioCharts.com


3) Regarding cash, I personally believe everyone should be realistic about it and account for a little bit in their allocation.  It's not just about maximizing every tiny percentage of possible return, but also about good planning and making your life easier.  I personally recommend for retirees to have 1 year of cash set aside no matter how they invest.  Top off your living expense fund at the end of the year (rebalancing in the most tax-efficient manner in the process) and leave your investments alone after that.  Assuming the standard 4% SWR, if you really think that having 1/25th of your portfolio in cash is going to kill your retirement chances then IMHO you're over-thinking it.

All that said, 100% stocks works fine for many people and I appreciate how you use VPW as an added safety measure and are realistic about your ability to make more money in the future.  If your primary goal for using VPW is to increase your withdrawal rate, I believe another way to approach the problem is to look at greater portfolio diversification beyond over-simplified "stocks" and "bonds".  But if 100% stocks is your thing and makes you most comfortable, your plan seems well researched and I think with that kind of thoughtfulness you'll do just fine.  Nice work!

[arebelspy]

2) Specifically referencing the charts you show from PortfolioCharts, the following comment jumped out: "As we all know, volatility in stock returns over 30 years gets very narrow…similar to bonds actually…just a bit further up the Y axis".  I think I know what you're getting at, but to be clear for those interpreting the charts: The uncertainty of the long-term average return gets lower the longer you hold an investment, but the annual volatility is exactly the same in year 30 as it is in year 1.  The distinction is very important for something like withdrawal rates.   Despite the 3% difference in CAGR (since 1972) for the two portfolios you show, the difference in long-term SWR over that same timeframe is less than 1%.  The high volatility of stocks negates much of the average returns benefit when considering drawdowns, and it's absolutely possible for a lower-return portfolio to be much safer in retirement than a higher-return option depending on the volatility.  Here are a few examples.  Note how the 60-40 and TSM (total stock market) pretty much match the chart above, and also how none of the top 3 SWR portfolios have more than 40% stocks.

(Emphasis added.)

This is very important to understand, and a bit counterintuitive at first. What makes it the case is the annual withdrawals.

If it were money sitting for decades, untouched, the higher CAGR should almost always win. But volatility matters a lot when making withdrawals, due to a higher chance of it being "down" when you have to make withdrawals.

I used to way undervalue volatility versus CAGR, but they both play an important role.

[Interest Compound]

The historical data show that the optimal asset allocation for multi-decade periods was always 100% stocks (or, depending on the precise settings for variables such as your investment expense ratio, near 100% stocks)--this allocation produced both the highest portfolio values and the highest success rates in every historical period covered by cfiresim.  So, if you assume that the future will be no worse than the past, then 100% stocks is (beyond argument) the best allocation.

Because I can already see Tyler (creator of PortfolioCharts.com) wincing at these statements, let me clarify that they, like the original post in this thread, were referring to asset allocation as a sliding scale solely between two options:  "stocks" and "bonds," each as a single unified asset class.  In reality, per Rebs' comment above regarding more diversified portfolios (like the permanent portfolio), there are of course specific asset allocations that have fared better historically once you start slicing and dicing these two generic asset classes into more specific subclasses and/or add other asset classes into the mix.  (My own favored approach continues to be to keep things simple and stick with 100% stocks, using a portfolio comprising broadly-diversified domestic and international total market index funds.)

Good to see you know me well.  ;)  Your comments are spot-on.  And your plan seems perfectly reasonable!

I appreciate the well-researched post, Interest Compound.  Just a few comments:

1) Yes, when you limit your investing options to a single broad stock fund and a broad bond fund, high percentages of stocks usually win out.  Interestingly, research has shown that 100% stocks has a lower SWR than lower stock allocations, and there's no real difference in retirement safety between 35% stocks and 80% stocks.  The sweet spot for maximizing both withdrawal rates and long-term returns for a traditional stock/bond portfolio seems to be about 75-80% stocks.  Importantly, other portfolios that use more diverse funds have different results.  (more on that below)

From Wade Pfau

Agreed, I'm totally with you on that. This type of research, and posts like this are what led me to my initial 80/20 allocation. VPW is what really flipped things around for me.

2) Specifically referencing the charts you show from PortfolioCharts, the following comment jumped out: "As we all know, volatility in stock returns over 30 years gets very narrow…similar to bonds actually…just a bit further up the Y axis".  I think I know what you're getting at, but to be clear for those interpreting the charts: The uncertainty of the long-term average return gets lower the longer you hold an investment, but the annual volatility is exactly the same in year 30 as it is in year 1.  The distinction is very important for something like withdrawal rates.   Despite the 3% difference in CAGR (since 1972) for the two portfolios you show, the difference in long-term SWR over that same timeframe is less than 1%.  The high volatility of stocks negates much of the average returns benefit when considering drawdowns, and it's absolutely possible for a lower-return portfolio to be much safer in retirement than a higher-return option depending on the volatility.  Here are a few examples.  Note how the 60-40 and TSM (total stock market) pretty much match the chart above, and also how none of the top 3 SWR portfolios have more than 40% stocks.

From PortfolioCharts.com

Totally with you on that too. This is the beauty of VPW. Volatility is bad, as withdrawals during a bad year can really kill the long-term growth of the portfolio (pretty much the opposite of the chart I showed with the "Huge Gap"). It seems VPW is the saving grace for the 100% stock portfolio, as it negates this. Here's what the VPW yearly withdrawals looked like starting in the same 1972, 40 year withdrawal period from your chart:

(Inflation-adjusted - Solid line is VPW - Dotted line is the 4% withdrawal rate)


It avoids crippling the portfolio in the starting years, then really explodes. The median withdrawal over the period is $71,000 and the average of the first 20 years is $50,700. This isn't an apples to apples comparison though, as VPW purposely draws down the portfolio to 0, while the 4% rule still has $2 million inflation-adjusted dollars for your heirs.

I had some trouble replicating the results from your chart, until I realized it was using 100% domestic stocks, whereas all my charts are cap weighting between domestic and international stocks (so 50/50). While 100% domestic stocks failed with a 5% withdrawal rate starting in 1972 (hitting 0 in 1994), the cap weight 50/50 domestic/international did better than the Permanent Portfolio:





Looking at VPW with 60/40, it doesn't really help too much during this time period. First I'll have to work a bit longer, as the initial withdrawal rate is 4.0% instead of the 5.1% of 100% stocks, then the median withdrawal moves down to $66,000. If I change the start year to something like 1950, it's even worse, with the median withdrawal dropping from $130k with 100% stocks, to $76k with 60/40 stocks/bonds.

To be honest, on principle I can't go with any of the alternative portfolios listed, permanent portfolio included. The allocations...20% here, 14% there, 4% here, they just seem like they were made to overfit the data. As my signature says, passive investing (as I define it) has a deep humility at its core--the aim is not to separate winners from losers, but rather to hold the entire market. Now you could argue I'm not staying true to this mantra, as this whole thread is about going 100% stocks (winners) and skipping out on Bonds (losers), but that's why I made this thread. So everyone here can tell me if I'm being an idiot :)

3) Regarding cash, I personally believe everyone should be realistic about it and account for a little bit in their allocation.  It's not just about maximizing every tiny percentage of possible return, but also about good planning and making your life easier.  I personally recommend for retirees to have 1 year of cash set aside no matter how they invest.  Top off your living expense fund at the end of the year (rebalancing in the most tax-efficient manner in the process) and leave your investments alone after that.  Assuming the standard 4% SWR, if you really think that having 1/25th of your portfolio in cash is going to kill your retirement chances then IMHO you're over-thinking it.

It's interesting you say that, because this is a staple of VPW, and it makes no sense to me. With VPW you keep 1 year of living expenses in cash, then at the end of the year you top it up with a withdrawal up to the recommended number. How does keeping 1 year of expenses in cash help you, when you're still making a withdrawal from your portfolio at the end of the year, regardless of market conditions? I'd think that taking the withdrawals monthly would be more efficient, since it keeps your money working for you longer, and most years the market goes up.

Can you explain your reasoning behind this?

All that said, 100% stocks works fine for many people and I appreciate how you use VPW as an added safety measure and are realistic about your ability to make more money in the future.  If your primary goal for using VPW is to increase your withdrawal rate, I believe another way to approach the problem is to look at greater portfolio diversification beyond over-simplified "stocks" and "bonds".  But if 100% stocks is your thing and makes you most comfortable, your plan seems well researched and I think with that kind of thoughtfulness you'll do just fine.  Nice work!

Thanks for your feedback Tyler! I haven't made the decision yet, but I'm close to pulling the trigger.

[Tyler]

I had some trouble replicating the results from your chart, until I realized it was using 100% domestic stocks, whereas all my charts are cap weighting between domestic and international stocks (so 50/50).  While 100% domestic stocks failed with a 5% withdrawal rate starting in 1972 (hitting 0 in 1994), the cap weight 50/50 domestic/international did better than the Permanent Portfolio:

Be careful not to look at only one starting year and extrapolate over every starting year.  For the tool you're using, re-run the same analysis for every start year (or use PC which does it for you ;) ). Looking at the single worst timeframes for each, the Permanent Portfolio has supported a higher WR than a 50/50 US/Int'l stock split using a traditional Trinity/Bengen study withdrawal system (I'll look into adding a VPW option).   Whether that means it's "better" for you depends on considering other factors and goals as well, and I have no problem with your 50/50 stock allocation.  I don't believe in a one-size-fits-all portfolio.

To be honest, on principle I can't go with any of the alternative portfolios listed, permanent portfolio included. The allocations...20% here, 14% there, 4% here, they just seem like they were made to overfit the data. As my signature says, passive investing (as I define it) has a deep humility at its core--the aim is not to separate winners from losers, but rather to hold the entire market. Now you could argue I'm not staying true to this mantra, as this whole thread is about going 100% stocks (winners) and skipping out on Bonds (losers), but that's why I made this thread. So everyone here can tell me if I'm being an idiot :)

I guess that's a matter of opinion.  The PP is equally split between four assets selected for specific economic conditions.  I really don't see that as over-fitting the data.  Same for something like Bernstein, Ivy, etc. with portfolios equally divided among several assets.  The Swedroe allocation may seem a little random, but the entire theory is built on investing in high-risk stocks so that you can get the same overall return with lower portfolio volatility while dialing back stock exposure.  Merriman may look over-fit until you read through his explanation, and then it makes sense.  Basically, I wouldn't judge the allocations of diverse portfolios until you take the time to read up and truly understand them.

One of the major themes of Modern Portfolio Theory is about thoughtfully selecting different assets that, when combined, have better overall performance than any individual asset on their own.  It's counter-intuitive when you only think about long-term averages, study the returns of each asset in isolation, or artificially limit yourself to two options.  But once you account for multiple diverse assets, correlations, and annual rebalancing it starts to make sense.  The William Bernstein books cover this topic pretty well. 

That said, I definitely do not recommend investing in any portfolio you do not understand or are not comfortable with for any reason.  If after considering the various tradeoffs a straightforward stock portfolio still makes the most sense to you, that's a perfectly fine plan. 

It's interesting you say that, because this is a staple of VPW, and it makes no sense to me. With VPW you keep 1 year of living expenses in cash, then at the end of the year you top it up with a withdrawal up to the recommended number. How does keeping 1 year of expenses in cash help you, when you're still making a withdrawal from your portfolio at the end of the year, regardless of market conditions? I'd think that taking the withdrawals monthly would be more efficient, since it keeps your money working for you longer, and most years the market goes up.

Can you explain your reasoning behind this?

My opinion is that having a year of cash is not about being most efficient economically but personally.  Most passive investment strategies call for simple annual rebalancing and leaving things alone the rest of the year.  Keeping that same schedule with withdrawals (rather than withdrawing money every month) is a lot less effort and supports the passive mindset.  IMHO, if you're trying to squeeze every last dime out of your investments by managing them monthly in retirement, you're not nearly as passive as you might think.  Keeping one year of expenses in cash should not have a significant effect on your overall portfolio performance (once you have 25+ saved up), but its benefit to your emotions and investing behavior are quite noticeable. 

When I first started FIRE, I tried only taking out a few months of cash at a time and frankly it was a hassle.  A year really does feel right, and the fact that doing it that way also allows me to easily calculate my expenses for the year and withdraw the money in the most tax-efficient way possible are also nice bonuses.  FWIW, my own portfolio includes a nice chunk of cash (on top of the annual expenses) as part of my AA. 

[Scandium]

Can you post a chart with the withdrawal amounts for these as well? That's my main issue, although you say it doesn't bother you so that's good. From your second post it looks like VPW drops from $40k/year to under $18k/year in the first 1/3 of retirement? That's a pretty huge drop in standard of living! I can imagine that being painful (even more so for my wife, and thus for me..). The chart shows it will get better after that, but I can imagine that if I had to spend 1/2 what I had planned for several years I might be scared that I had screwed up, and start looking for work. Had I been out of the workforce a while, or I was old, and had trouble getting a job it might start to really panic.

Some variability in withdraws is fine, and probably inevitable, but cutting costs in 1/2 might seem nigh near impossible, and not sure it's something I'd like to live through to gain 1% in SWR.. Personal preference though.

[Interest Compound]

Can you post a chart with the withdrawal amounts for these as well? That's my main issue, although you say it doesn't bother you so that's good. From your second post it looks like VPW drops from $40k/year to under $18k/year in the first 1/3 of retirement? That's a pretty huge drop in standard of living! I can imagine that being painful (even more so for my wife, and thus for me..). The chart shows it will get better after that, but I can imagine that if I had to spend 1/2 what I had planned for several years I might be scared that I had screwed up, and start looking for work. Had I been out of the workforce a while, or I was old, and had trouble getting a job it might start to really panic.

Some variability in withdraws is fine, and probably inevitable, but cutting costs in 1/2 might seem nigh near impossible, and not sure it's something I'd like to live through to gain 1% in SWR.. Personal preference though.

Remember, I'm making a big distinction here between retirement expense needs, and withdrawal amount. Also, due to a lack of long-term data for international stocks, these two charts are 100% domestic stocks.

That $18k a year number is derived from looking at the worst case historical scenarios, as a 30 year old, who lives off their portfolio until age 100. Those same worst case scenarios would've resulted in a portfolio wipe with the 4% rule:



In the worst case scenario of 1929, VPW would have you withdrawing less than you want for 23 of the first 26 years. The remaining 44 years have an average withdrawal of $73,800, while your 4% rule counterpart lost their ability to feed themselves after 22 years.



Similar story for the 1966 investor, where VPW would have you withdrawing less for 16 of the first 21 years, while the remaining years (so far) would have you averaging a $59,000 withdrawal. While your 4% rule counterpart lost their ability to feed themselves after 25 years.



If you use 50/50 domestic/international stocks for the 1966 case (international data is available starting in 1972, so only the first few years are 100% domestic) the 4% rule lasts a bit longer, but now you end up broke at age 77...I'm not sure if that's better or worse:



The typical scenario for VPW looks more like this:



The worst case scenarios for VPW have you spending less for about two decades so the portfolio can recover for the rest of your life, while those same years for the 4% rule would have resulted in portfolio wipes. In short, either way you end up working again. VPW has you doing it while you're younger, on a part-time basis doing the things you love (hopefully!), with $700,000 in the bank. The 4% rule has you doing it when you're 55, on a full-time basis, with $0 in the bank.

Pick your poison :-P

[Interest Compound]

I had some trouble replicating the results from your chart, until I realized it was using 100% domestic stocks, whereas all my charts are cap weighting between domestic and international stocks (so 50/50).  While 100% domestic stocks failed with a 5% withdrawal rate starting in 1972 (hitting 0 in 1994), the cap weight 50/50 domestic/international did better than the Permanent Portfolio:

Be careful not to look at only one starting year and extrapolate over every starting year.  For the tool you're using, re-run the same analysis for every start year (or use PC which does it for you ;) ). Looking at the single worst timeframes for each, the Permanent Portfolio has supported a higher WR than a 50/50 US/Int'l stock split using a traditional Trinity/Bengen study withdrawal system (I'll look into adding a VPW option).   Whether that means it's "better" for you depends on considering other factors and goals as well, and I have no problem with your 50/50 stock allocation.  I don't believe in a one-size-fits-all portfolio.

To be honest, on principle I can't go with any of the alternative portfolios listed, permanent portfolio included. The allocations...20% here, 14% there, 4% here, they just seem like they were made to overfit the data. As my signature says, passive investing (as I define it) has a deep humility at its core--the aim is not to separate winners from losers, but rather to hold the entire market. Now you could argue I'm not staying true to this mantra, as this whole thread is about going 100% stocks (winners) and skipping out on Bonds (losers), but that's why I made this thread. So everyone here can tell me if I'm being an idiot :)

I guess that's a matter of opinion.  The PP is equally split between four assets selected for specific economic conditions.  I really don't see that as over-fitting the data.  Same for something like Bernstein, Ivy, etc. with portfolios equally divided among several assets.  The Swedroe allocation may seem a little random, but the entire theory is built on investing in high-risk stocks so that you can get the same overall return with lower portfolio volatility while dialing back stock exposure.  Merriman may look over-fit until you read through his explanation, and then it makes sense.  Basically, I wouldn't judge the allocations of diverse portfolios until you take the time to read up and truly understand them.

One of the major themes of Modern Portfolio Theory is about thoughtfully selecting different assets that, when combined, have better overall performance than any individual asset on their own.  It's counter-intuitive when you only think about long-term averages, study the returns of each asset in isolation, or artificially limit yourself to two options.  But once you account for multiple diverse assets, correlations, and annual rebalancing it starts to make sense.  The William Bernstein books cover this topic pretty well. 

That said, I definitely do not recommend investing in any portfolio you do not understand or are not comfortable with for any reason.  If after considering the various tradeoffs a straightforward stock portfolio still makes the most sense to you, that's a perfectly fine plan. 

It's interesting you say that, because this is a staple of VPW, and it makes no sense to me. With VPW you keep 1 year of living expenses in cash, then at the end of the year you top it up with a withdrawal up to the recommended number. How does keeping 1 year of expenses in cash help you, when you're still making a withdrawal from your portfolio at the end of the year, regardless of market conditions? I'd think that taking the withdrawals monthly would be more efficient, since it keeps your money working for you longer, and most years the market goes up.

Can you explain your reasoning behind this?

My opinion is that having a year of cash is not about being most efficient economically but personally.  Most passive investment strategies call for simple annual rebalancing and leaving things alone the rest of the year.  Keeping that same schedule with withdrawals (rather than withdrawing money every month) is a lot less effort and supports the passive mindset.  IMHO, if you're trying to squeeze every last dime out of your investments by managing them monthly in retirement, you're not nearly as passive as you might think.  Keeping one year of expenses in cash should not have a significant effect on your overall portfolio performance (once you have 25+ saved up), but its benefit to your emotions and investing behavior are quite noticeable. 

When I first started FIRE, I tried only taking out a few months of cash at a time and frankly it was a hassle.  A year really does feel right, and the fact that doing it that way also allows me to easily calculate my expenses for the year and withdraw the money in the most tax-efficient way possible are also nice bonuses.  FWIW, my own portfolio includes a nice chunk of cash (on top of the annual expenses) as part of my AA.

I think it might be best to leave this thread to only stocks/bonds, as I need to wrap my mind around your comments on the other portfolios. I'll probably make a new thread in a week or two on that :)

Regarding the year of cash, I always assumed I'd simply setup an automatic monthly withdrawal from my Vanguard account at the beginning of the year, call it my "paycheck", and leave it alone for the year. After typing that, and re-reading your post, I can already see some things I'd want to change.

"Oh, I made enough money with my hobbies last month to cover next month's expenses, let's cancel that Vanguard paycheck for next month!"

"Oh, we've had a big crash, let's recalculate VPW for the year and reset the monthly withdrawals!"

Hmm...having a year of buffer in cash makes a lot of sense.

Thanks for the insight Tyler!

[Interest Compound]

Can you post a chart with the withdrawal amounts for these as well? That's my main issue, although you say it doesn't bother you so that's good. From your second post it looks like VPW drops from $40k/year to under $18k/year in the first 1/3 of retirement? That's a pretty huge drop in standard of living! I can imagine that being painful (even more so for my wife, and thus for me..). The chart shows it will get better after that, but I can imagine that if I had to spend 1/2 what I had planned for several years I might be scared that I had screwed up, and start looking for work. Had I been out of the workforce a while, or I was old, and had trouble getting a job it might start to really panic.

Some variability in withdraws is fine, and probably inevitable, but cutting costs in 1/2 might seem nigh near impossible, and not sure it's something I'd like to live through to gain 1% in SWR.. Personal preference though.

Some additional comments in addition to the charts from earlier.

VPW is a rational solution to the SWR dillema: Markets went down a lot so should I reduce my withdrawals? By how much? Markets went up a lot so should I increase my withdrawals? By how much? VPW provides a mathematical answer that will protect the portfolio from ruin, yet keep targeting portfolio depletion on an exact selected year.

VPW is not a unique method. The principles on which it is based (e.g. using an amortization schedule to determine withdrawal rates) predate it. The key new contribution of VPW is that it adapts withdrawal rates to the asset allocation, it allows for a flexible last-withdrawal age, and it comes with a neat backtesting spreadsheet to evaluate its historical behavior.

[arebelspy]

The typical scenario for VPW looks more like this:

Those huge dives in 2000, 2007, etc.  Ugly.  Having to "cut back" doesn't sound appealing to me, I'd much prefer stability. Something similar to the below (which I just made up and eyeballed, but the idea of raising steadily in real terms when the portfolio supports it):

[Interest Compound]

The typical scenario for VPW looks more like this:

Those huge dives in 2000, 2007, etc.  Ugly.  Having to "cut back" doesn't sound appealing to me, I'd much prefer stability. Something similar to the below (which I just made up and eyeballed, but the idea of raising steadily in real terms when the portfolio supports it):

My actual spending will probably mirror your blue line. Well, I'm sure there will be some spike years when the kids go to college, or some emergency/unexpected expense happens, but that's close enough. But why should my paycheck be at all related to my spending? Remember, I'm making a distinction here between retirement spending, and withdrawal amount.

I don't want to live paycheck to paycheck in retirement and I view VPW as a paycheck. In other words, just because VPW says I can withdrawal $83,000 one year, doesn't mean I'm going to go out and find something to spend an extra $43,000 on that year! :) The only thing that should be cut back in 2000 and 2007, is your paycheck. Unless you were living paycheck to paycheck, and immediately grew your lifestyle with your paycheck, you shouldn't feel a thing.

Seriously, if you were offered a 50% raise during your working years, would you turn it down because it had "a lot of variability"? Of course not. You'd take the money and either spend it, pad your emergency fund, or invest it. Since we are all mustachians here, we'd likely end up saving/investing most of it.

From what I've seen, there are two ways people handle this with VPW:

1. They literally treat it like a paycheck, withdrawal every year on Jan 1st, and leave whatever they don't spend growing in their savings account for emergencies. This is akin to someone working, saving maybe 30% of their "paycheck", and putting it away for a rainy day.

2. They literally treat it like a paycheck, withdrawal every year on Jan 1st, but they only withdrawal enough to top off their 1 year buffer of cash. This is akin to someone working, saving maybe 30% of their "paycheck", and investing it.

The person following #1 will have a steadily growing allocation to cash. You could make the argument they're no longer "100% stocks", but they usually don't see it that way. They see it as their emergency/vacation/spending account. The person following #2 will build a bigger and bigger buffer for their withdrawals in the following years. If the person who retired in 1955 from our chart, only spends $40,000 a year for the first 5 years (exactly what their expenses were before retirement), and invests the rest (following #2's plan), they will have earned themselves a 10% raise for the rest of their lives. By this I mean their VPW withdrawal each year will have increased by 10%. If they do this for the first 10 years, they will have earned themselves a 30% raise for the rest of their lives.

We've got a lot of folks here preparing for the worst with a 4% SWR, lowering their retirement paycheck, putting money in bonds to reduce paycheck volatility, and living paycheck to paycheck. I think I've found a better way.

[arebelspy]

Withdrawing more than you are going to spend seems silly to me.

Actual withdrawal amount should match spending (even if potential withdrawal amount is much higher).  Thus the blue line.

[Rubic]

Withdrawing more than you are going to spend seems silly to me.

One possible scenario where it might make sense would be to withdraw funds from a taxable account where the long-term capital gains tax would be zero for the current year (e.g. $37K withdrawal on a budget of $28K expenses).  In the following year you could start paying your living expenses out of these funds until you clarified what your tax situation would be for the next year.

Even if you decided to later re-invest the funds, you would have still reset your cost basis to lower your taxes for withdrawals in future years.

[arebelspy]

Withdrawing more than you are going to spend seems silly to me.

One possible scenario where it might make sense would be to withdraw funds from a taxable account where the long-term capital gains tax would be zero for the current year (e.g. $37K withdrawal on a budget of $28K expenses).  In the following year you could start paying your living expenses out of these funds until you clarified what your tax situation would be for the next year.

Even if you decided to later re-invest the funds, you would have still reset your cost basis to lower your taxes for withdrawals in future years.

IMO you'd be better off doing a Roth rollover in that case.

[Rubic]

Withdrawing more than you are going to spend seems silly to me.

One possible scenario where it might make sense would be to withdraw funds from a taxable account where the long-term capital gains tax would be zero for the current year (e.g. $37K withdrawal on a budget of $28K expenses).  In the following year you could start paying your living expenses out of these funds until you clarified what your tax situation would be for the next year.

Even if you decided to later re-invest the funds, you would have still reset your cost basis to lower your taxes for withdrawals in future years.

IMO you'd be better off doing a Roth rollover in that case.

Agreed, if you have any remaining Roth conversions available at that time.

[arebelspy]

Withdrawing more than you are going to spend seems silly to me.

One possible scenario where it might make sense would be to withdraw funds from a taxable account where the long-term capital gains tax would be zero for the current year (e.g. $37K withdrawal on a budget of $28K expenses).  In the following year you could start paying your living expenses out of these funds until you clarified what your tax situation would be for the next year.

Even if you decided to later re-invest the funds, you would have still reset your cost basis to lower your taxes for withdrawals in future years.

IMO you'd be better off doing a Roth rollover in that case.

Agreed, if you have any remaining Roth conversions available at that time.

Is there a scenario where you couldn't backdoor Roth, if you're in a 0% LTCG bracket?

[Interest Compound]

From what I've seen, there are two ways people handle this with VPW:

1. They literally treat it like a paycheck, withdrawal every year on Jan 1st, and leave whatever they don't spend growing in their savings account for emergencies. This is akin to someone working, saving maybe 30% of their "paycheck", and putting it away for a rainy day.

2. They literally treat it like a paycheck, withdrawal every year on Jan 1st, but they only withdrawal enough to top off their 1 year buffer of cash. This is akin to someone working, saving maybe 30% of their "paycheck", and investing it.

The person following #1 will have a steadily growing allocation to cash. You could make the argument they're no longer "100% stocks", but they usually don't see it that way. They see it as their emergency/vacation/spending account. The person following #2 will build a bigger and bigger buffer for their withdrawals in the following years. If the person who retired in 1955 from our chart, only spends $40,000 a year for the first 5 years (exactly what their expenses were before retirement), and invests the rest (following #2's plan), they will have earned themselves a 10% raise for the rest of their lives. By this I mean their VPW withdrawal each year will have increased by 10%. If they do this for the first 10 years, they will have earned themselves a 30% raise for the rest of their lives.

I've thought about this a bit more. The red line is person #1's paycheck throughout their retirement. They can choose to spend it, save it, whatever. The red line is essentially their new job, and they may do whatever they wish with their paychecks. The blue line is person #2's withdrawn paycheck. I went through each year, looked at the VPW paycheck, and decided how much person #2 needed to spend that year. I modeled some surprise medical expenses, college payments, buying a new car...etc. I never once went above my VPW paycheck, and each year I "saved" about 20-50% of it. There were a few years during the mid 70's and 80's where I took out the max, but those were exceptions.

I think person #1's strategy can make sense. Think about it as rebalancing to cash, but only after you've met your expenses for the year, and any money in the cash allocation can be spent anytime without worry of depleting your portfolio. I don't think I'd take that path, but it makes sense to me. I can see how someone would prefer that level of safety.

[arebelspy]

Withdrawing more than you're going to spend and leaving it sitting in cash, rather than invested, sounds like a terrible idea, and a good way to hurt your portfolio's longevity chances.

[Rubic]

Withdrawing more than you are going to spend seems silly to me.

One possible scenario where it might make sense would be to withdraw funds from a taxable account where the long-term capital gains tax would be zero for the current year (e.g. $37K withdrawal on a budget of $28K expenses).  In the following year you could start paying your living expenses out of these funds until you clarified what your tax situation would be for the next year.

Even if you decided to later re-invest the funds, you would have still reset your cost basis to lower your taxes for withdrawals in future years.

IMO you'd be better off doing a Roth rollover in that case.

Agreed, if you have any remaining Roth conversions available at that time.

Is there a scenario where you couldn't backdoor Roth, if you're in a 0% LTCG bracket?

Perhaps I'm not understanding the advantage of the Backdoor Roth in this situation.  From http://thefinancebuff.com/the-backdoor-roth-ira-a-complete-how-to.html, my understanding is that it's useful if my income exceeds the contribution limit.

In the scenario I'm describing, traditional IRA funds have been converted to a Roth over a period of years to minimize the effects of RMD at age 70.  I would have funds remaining in taxable accounts and could withdraw $40K+ per year at zero long-term capital gains tax.  If there are sufficient funds and my expenses are below $40K/year, I'm living essentially tax-free until SS benefits are withdrawn (and taxed) at age 70.

[Interest Compound]

Withdrawing more than you're going to spend and leaving it sitting in cash, rather than invested, sounds like a terrible idea, and a good way to hurt your portfolio's longevity chances.

That's how I feel too, and maybe with the 4% rule it's true. But remember, all the VPW graphs I've shown thus far, including the worst-case scenarios, are following person #1's path, as if they literally threw away their cash allocation each year. Even with this, the portfolio's longevity surpassed the 4% rule, with a higher median withdrawal. If someone feels safer keeping their excess in cash during retirement, it won't hurt them too bad. Heck, they can even put it in CDs if they wanted to, and feel just as safe. Then it would look like your typical retirement glide-path :-P

But let's not get too hung up on this. I agree person #2's path is better, and it's the one I'd choose.

[Kaspian]

The math for this works for me 100% of the time.  You can't create pie charts or graphs for emotion and that is the grenade pin as to why this allocation ends up failing for the majority of investors.  Honestly--I wholeheartedly agree with all of the data presented but where are the psychological formulas?  (Us programmers, engineers, accountants, etc., are good at creating efficient systems, but I've been in enough relationships to know my systems fail when presented with human reality.) 

[brooklynguy]

Even with this, the portfolio's longevity surpassed the 4% rule, with a higher median withdrawal.

As you noted in the other VPW thread, it's difficult to use any single metric to compare different withdrawal strategies because it will always focus on only one of the multiple tradeoffs involved (for example, the traditional "success rate" metric focuses exclusively on likelihood of avoiding total portfolio depletion but otherwise ignores variability in terminal portfolio value and completely ignores spending levels throughout the retirement period).

Have you seen the comparison of different variable withdrawal strategies that Pfau did?  It's available here:  "Making Sense Out of Variable Spending Strategies for Retirees" (and was discussed in this thread:  Pfau on Variable Withdrawal Strategies).  To do an apples-to-apples performance comparison of the various withdrawal strategies, he used an "XYZ" benchmark for all of them -- the X% probability that spending falls below a threshold of $Y (in inflation-adjusted dollars) by year Z of retirement.  As noted in the linked thread, his methodology used Monte Carlo simulations with built-in capital market assumptions and therefore produced what many of us view as overly pessimistic results across the board for all of the strategies, but I thought the XYZ metric was an interesting way of comparing different withdrawal strategies (both variable and non-variable) against one another.

[Tyler]

The whole concept of the MMM style of early retirement is built around first reducing your expenses to the minimum, happy, sustainable level and then building your finances and retirement plan around that.  For me, the traditional fixed purchasing power assumption for withdrawal rates best matches that mindset.  My spending has been X for years now, and I'm very happy with that.  So I'll plan my retirement around a relatively constant X.  Of course I'll be smart and flexible about it, but firing up a spreadsheet that tells me I can can now spend way more or way less just seems kinda unnecessary and counter to everything I've built. 

The problem I think I have with VPW (and other variable withdrawal strategies) is that they work the other direction, often looking to maximize spending without regard to how much is actually required to be happy.  Or focusing so much on the not working part that they pretend drastically reducing spending below a happy level will be just fine and stress free.  It does strike me as a good method for non-Mustachinan retirees who love to spend money when they can.  But I see little benefit for me personally. 

[Pooplips]

Comment to follow

[steveo]

The problem I think I have with VPW (and other variable withdrawal strategies) is that they work the other direction, often looking to maximize spending without regard to how much is actually required to be happy.  Or focusing so much on the not working part that they pretend drastically reducing spending below a happy level will be just fine and stress free.  It does strike me as a good method for non-Mustachinan retirees who love to spend money when they can.  But I see little benefit for me personally.

This is a good point. If I choose to use a VPW it would be to save the additional cash when returns were great and use that money instead of withdrawing when returns were poor. I don't see the need to increase spending and I don't think its possible to decrease spending significantly.

Still I think that there is something to this approach. Withdraw more when returns are great and place the money into safer assets. When returns are poor withdraw from the safer assets. Personally I would do this anyway but you can't model this so well in a statistical model.

[Retire-Canada]

If VPW tells you your allowed withdrawal amount is $70K and you only need $40K you can just leave the "excess" $30K in the portfolio.

When you calculate the following year's withdrawal amount the that $30K will be indistinguishable from having a higher return over the last year.

[arebelspy]

The whole concept of the MMM style of early retirement is built around first reducing your expenses to the minimum, happy, sustainable level and then building your finances and retirement plan around that.  For me, the traditional fixed purchasing power assumption for withdrawal rates best matches that mindset.  My spending has been X for years now, and I'm very happy with that.  So I'll plan my retirement around a relatively constant X.  Of course I'll be smart and flexible about it, but firing up a spreadsheet that tells me I can can now spend way more or way less just seems kinda unnecessary and counter to everything I've built. 

The problem I think I have with VPW (and other variable withdrawal strategies) is that they work the other direction, often looking to maximize spending without regard to how much is actually required to be happy.  Or focusing so much on the not working part that they pretend drastically reducing spending below a happy level will be just fine and stress free.  It does strike me as a good method for non-Mustachinan retirees who love to spend money when they can.  But I see little benefit for me personally.

Good point. The higher spending concept didn't bother me as much as the whiplash from lower spending simply because I, like you, have spent around X, and, being as optimized as I can, decreasing expenses a significant amount would impact my happiness.  So I don't want to decrease expenses, and VPW gives a very significant chance of that happening over the 4% WR (which, if strictly followed, never decreases expenses).

The idea that higher expenses won't make me happier is also true.  So VPW offers only downside (less happiness via significantly decreased expenses) and no upside (no more happiness via higher expenses).  For me, at least.

The problem I think I have with VPW (and other variable withdrawal strategies) is that they work the other direction, often looking to maximize spending without regard to how much is actually required to be happy.  Or focusing so much on the not working part that they pretend drastically reducing spending below a happy level will be just fine and stress free.  It does strike me as a good method for non-Mustachinan retirees who love to spend money when they can.  But I see little benefit for me personally.

This is a good point. If I choose to use a VPW it would be to save the additional cash when returns were great and use that money instead of withdrawing when returns were poor. I don't see the need to increase spending and I don't think its possible to decrease spending significantly.

The problem is, there is no "additional cash."  Your returns are the same regardless of your withdrawal method.  So by taking out more from your AA into a lower returning asset when returns are high, you are decreasing the chance of your portfolio lasting.

Still I think that there is something to this approach. Withdraw more when returns are great and place the money into safer assets. When returns are poor withdraw from the safer assets. Personally I would do this anyway but you can't model this so well in a statistical model.

If you withdraw more when returns are great, you will have to decrease spending when returns are poor, as you won't have the assets returning enough to save you.  If you don't decrease spending, yes, you can use some of that money you withdrew and didn't spend, until it runs out, but it would have been better to not withdraw it in the first place.

If VPW tells you your allowed withdrawal amount is $70K and you only need $40K you can just leave the "excess" $30K in the portfolio.

When you calculate the following year's withdrawal amount the that $30K will be indistinguishable from having a higher return over the last year.

Naturally.  But then what's the point of using VPW if you don't actually use it?  The best I can figure is to figure out a max withdrawal number for that year.  But if you're just going to take out your same amount anyways, and let your portfolio grow by the excess (the smart thing to do anyways, IMO, increasing your chance of portfolio longevity), you aren't really using VPW, you're using a more normal X% rule.

[Radagast]

Naturally.  But then what's the point of using VPW if you don't actually use it?  The best I can figure is to figure out a max withdrawal number for that year.  But if you're just going to take out your same amount anyways, and let your portfolio grow by the excess (the smart thing to do anyways, IMO, increasing your chance of portfolio longevity), you aren't really using VPW, you're using a more normal X% rule.

You could be like Bernstein and put it in I-bonds or TIPS, or mix in some other bonds, you could even stick it in a PP. Sort of like taking your winnings off the table. Or you can artificially force your returns to be inflation +1% after a 4% withdrawal every year regardless of market performance. That might be too much work for the benefit, but I am reading "Value Averaging" right now (does that sound like an exciting read or what).

[steveo]

The problem is, there is no "additional cash."  Your returns are the same regardless of your withdrawal method.  So by taking out more from your AA into a lower returning asset when returns are high, you are decreasing the chance of your portfolio lasting.

I'm not sold on this point. Its a little similar to rebalancing and I think there is a chance that it could increase your returns. You cash out a little during the good times and have some buffer during the bad times.

[arebelspy]

The problem is, there is no "additional cash."  Your returns are the same regardless of your withdrawal method.  So by taking out more from your AA into a lower returning asset when returns are high, you are decreasing the chance of your portfolio lasting.

I'm not sold on this point. Its a little similar to rebalancing and I think there is a chance that it could increase your returns. You cash out a little during the good times and have some buffer during the bad times.

Okay.  Explain to me how VPW generates additional money (looking at overall net worth).

[arebelspy]

Tyler wrote a * awesome post on this topic (the main topic of the thread, 100% stock allocation, not the VPW, which we've sidetracked to) yesterday:
http://portfoliocharts.com/2015/11/17/how-safe-withdrawal-rates-work/

[Retire-Canada]

Naturally.  But then what's the point of using VPW if you don't actually use it?  The best I can figure is to figure out a max withdrawal number for that year.  But if you're just going to take out your same amount anyways, and let your portfolio grow by the excess (the smart thing to do anyways, IMO, increasing your chance of portfolio longevity), you aren't really using VPW, you're using a more normal X% rule.

So in my own case I have 2 budgets: #1 for essential or mostly essential items that I would like to spend each year [phone, mortgage, food, etc..] and #2 for non-essentials [new sports gear, international travel, larger home maintenance projects, new computer, etc...] that I don't need in any given year, but over say 10yrs I'll want to spend $$ in each area. I can defer those costs for a period of time if my investments are not doing well.

So my annual spending would not be the same each year if my portfolio allowed it. I've used the variable spending options in cFIREsim to see what that looks like and if I try a $40K/yr fixed WR vs a $30-$50K VWR I get similar backtesting success rates for much lower saved starting amounts. It would make the difference between working a number of extra years if I wanted to stick with the fixed 4% WR.

I haven't decided on exactly how I will determine when to spend $30K and when to spend $50K. I could use VPW to give me this signal as well as a signal when I needed to bring in additional income if it was suggesting a WR amount less than my base costs.

Now if it started to tell me I could take out $200K/yr I have no useful way to spend that kind of $$ so I would spend my base amount and pay for any items on my non-essential list I wanted to deal with in a given year. Then leave the rest to compound.

[brooklynguy]

Tyler wrote a * awesome post on this topic (the main topic of the thread, 100% stock allocation, not the VPW, which we've sidetracked to) yesterday:
http://portfoliocharts.com/2015/11/17/how-safe-withdrawal-rates-work/

Tyler's post is awesome.  In my view, it reinforces the notion that, solely as between "stocks" and "bonds," 100% stocks is the best option for long retirements.  As an initial matter, I would note that the data used in his analysis only goes back to 1972, which means that the sample size of long retirement periods is quite limited.  Despite that, the data supports the same arguments made in the old forum threads and the Go Curry Cracker post Interest Compound cross-referenced in the OP (which all relied on the much bigger data set contained in cFIREsim) -- that high stock allocations had the highest historical success rates, with negligible differences in success rates among the various degrees of "high" stock allocation, but very meaningful differences in terminal portfolio values.  So 100% stocks historically had a success rate on par with the highest success rate of any stock/bond allocation whatsoever, but significantly higher average terminal portfolio values.  As always, though, Tyler makes a compelling case for looking beyond the generic total stock market and total bond market alternatives for additional portfolio fine-tuning options.

[arebelspy]

As always, though, Tyler makes a compelling case for looking beyond the generic total stock market and total bond market alternatives for additional portfolio fine-tuning options.

Indeed. But how to avoid the curve fitting from data mining?

The recent post on the Merriman Ultimate Buy and Hold Portfolio seems very much to be a case of this, to me.

That's one plus for the PP in my mind--1/4th to each of 4 asset types, done.  No "well 17.35% of this one and 29.84% of this one, etc. performed best historically."  Not that I'm a huge PP fan, but at least I know it isn't curve fitting.

[brooklynguy]

Indeed. But how to avoid the curve fitting from data mining?

The recent post on the Merriman Ultimate Buy and Hold Portfolio seems very much to be a case of this, to me.

That's one plus for the PP in my mind--1/4th to each of 4 asset types, done.  No "well 17.35% of this one and 29.84% of this one, etc. performed best historically."  Not that I'm a huge PP fan, but at least I know it isn't curve fitting.

Yep, that's a big concern for any highly customized portfolio.  I think you need to view any highly specific allocation with an extra grain of salt.  Tyler's response in post # 11 above (that once you read the explanation for some of these recommended portfolios, then they start to make sense) only gets you so far, because one can come up with a reasonable sounding explanation after the fact for almost any historically high performing allocation derived from backtesting.

[arebelspy]

because one can come up with a reasonable sounding explanation after the fact for almost any historically high performing allocation derived from backtesting.

Not always; I didn't think the Merriman explanation was particularly compelling, and I still haven't heard a reasonable sounding explanation for dual momentum.  ;)

But I agree, it's tough to find something that works and be convinced as to why it works, without some post-hoc fallacies potentially going on.

[Aphalite]

Not always; I didn't think the Merriman explanation was particularly compelling, and I still haven't heard a reasonable sounding explanation for dual momentum.  ;)

But I agree, it's tough to find something that works and be convinced as to why it works, without some post-hoc fallacies potentially going on.

Regarding Merriman - anyone who uses standard deviation as "risk" is just vomit-inducing, in my opinion. And I know you're joking about dual momentum, but I do think that the theory makes a lot of sense - get out when valuation is too crazy, and buy the best performing sectors since they tend to go up more so than the total market in the short term - BUT, in reality, it's a lot harder to pull off real time, and if you're going to use some sort of active strategy, focusing on fundamentals, imo, is a much better method than looking at charts. Plus, it makes you smarter and more in tune with what you actually own - I think REIs have a huge advantage in this regard vs index investors - they (for the most part) know exactly what they own and are intimate with the risks and what could go wrong

I think the whole obsession with finding a good strategy and how things backtests is a testament to 1) how poorly the human brain judges probabilities of outcomes, and 2) just how much people, no matter how intelligent, looks for shortcuts. The best strategy involves a large time commitment, and poring over regulatory filings isn't exactly 99% of the population's idea of fun. But looking for shortcuts to that has its downsides - when you introduce needless complexity like specifically tailored and backtested asset allocations in an effort to "optimize", you're participating in what Charlie Munger refers to as "Physics Envy". Trying to fit real life into a fancy equation is what leads to people like Merriman depending on standard deviation and beta as a measure of "risk". If you want to avoid sinking hours into studying businesses, then buy a total market US fund and just leave it alone.

[Tyler]

Great points about the limits of backtesting and dangers of over-fitting.  Just because a particular portfolio looks good looking backwards or the originator can tell an eloquent story about it doesn't mean it's a good idea today.  On the other hand, just because a portfolio has surprising returns doesn't mean we should automatically dismiss it as an over-fitted portfolio.  Discerning the difference takes a lot of wisdom.  Come to think of it, that deserves its own future post.  Thanks!

The Merriman portfolio sorta treads that line for me, and maybe wasn't the best example to reference earlier in the thread.  Regardless of the reasoning, however, I do appreciate that the end result is well-diversified so I definitely don't think it's a bad idea (or I wouldn't have referenced it at all). 

I agree that the PP is a good example of a portfolio that originated from something other than data mining and curve fitting.  Even if after reading up on the theory behind it you don't personally use the exact assets Harry Browne landed on, following his advice and thinking will generally point you towards an excellent asset allocation built on a solid foundation. 

[arebelspy]

Not always; I didn't think the Merriman explanation was particularly compelling, and I still haven't heard a reasonable sounding explanation for dual momentum.  ;)

But I agree, it's tough to find something that works and be convinced as to why it works, without some post-hoc fallacies potentially going on.

Regarding Merriman - anyone who uses standard deviation as "risk" is just vomit-inducing, in my opinion. And I know you're joking about dual momentum, but I do think that the theory makes a lot of sense - get out when valuation is too crazy, and buy the best performing sectors since they tend to go up more so than the total market in the short term - BUT, in reality, it's a lot harder to pull off real time

Of course market timing is great, when it's successful.   No one's disputing that.  That just may not happen as regularly as active investors hope.  :)

[AdrianC]

The math for this works for me 100% of the time.  You can't create pie charts or graphs for emotion and that is the grenade pin as to why this allocation ends up failing for the majority of investors.  Honestly--I wholeheartedly agree with all of the data presented but where are the psychological formulas?  (Us programmers, engineers, accountants, etc., are good at creating efficient systems, but I've been in enough relationships to know my systems fail when presented with human reality.)

The math works.

For us:
Investments: Stocks and a few I-bonds bought many years ago at good interest rates.

Anchor: Paid for house, zero debt and some cash and a willingness to not pull money out of the investment portion until absolutely necessary.

We don't need bonds. The house is the security blanket.

IMHO, stocks will not give the returns many people are expecting. Bogle says returns might be something like 4% annually before inflation over the next ten years, and I think he could be right. And bonds will be worse than that.

[frugalnacho]

I don't really get it.  Of course a variable strategy that is designed to vary the amount you withdraw to ensure it can't fail, will not fail by definition, because it can't, you simply lower the amount to withdraw until it guarantees success.  Why can't you just do that with a modified 4% strategy?  Like plan on 40k/yr, but then when the shit hits the fan you adjust down to 18k/yr? Isn't that essentially going to do the same thing?  The problem is that if I could live on only 18k/yr, then I worked way too long and seriously over saved if I have a stash that generates 40k/yr.

[CorpRaider]

Bogle's comment is focused on just US stocks/assets though, EM and EAFE are cheap and have higher forward expected returns based on all the jazz Bogle and others look at to predict that stuff.

[Retire-Canada]

I don't really get it.  Of course a variable strategy that is designed to vary the amount you withdraw to ensure it can't fail, will not fail by definition, because it can't, you simply lower the amount to withdraw until it guarantees success.  Why can't you just do that with a modified 4% strategy?  Like plan on 40k/yr, but then when the shit hits the fan you adjust down to 18k/yr? Isn't that essentially going to do the same thing?  The problem is that if I could live on only 18k/yr, then I worked way too long and seriously over saved if I have a stash that generates 40k/yr.

You can definitely use some other variable withdrawal strategy. You can go by gut feel or by specific signals. VPW could provide the specific signals if you cared to use it for that purpose.

I guess one thing I don't understand about the talk of steady COL withdrawals is that my expenses include a component that is not tied to a specific year, but over the course of a decade say I would need to address. So it would be quite possible to go from say $40K/yr avg WR to $25K for a period if shit hit the fan with my investments, but that hardly means that $25K/yr is all I need to live off and I over saved to hit $40K/yr WR on average.

Your roof has a few leaks and your investments have tanked so you patch the leaks and replace the roof a few years later when things recover.

Your SO wants to renovate the kitchen and your investments have tanked so you just paint and do the more elaborate upgrades later.

I could easily reduce my COL by 25%+ for a year or two, but at some point I'd have to expend the deferred costs.

If you look at the amount you have to save for a fixed WR vs a variable WR with floor and ceiling amounts you could say that by not being flexible you have to over save to pay for the lack of flexibility.