Stop worrying about the 4% rule

[farfromfire]

Specific words, sentences and ideas in a post can be critiqued, especially when they describe separate ideas. Or is one only allowed to respond to the last sentence?

[mathjak107 is a kinder soul than I and responded to the point. In this forum, both glides and reverse glides for post-retirement have been discussed in great detail, dismissing the "dumbest thing in the world" to Strick]

Though bananas are tasty, banana-flavored food is often not tasty at all.

[Monkey Uncle]

Specific words, sentences and ideas in a post can be critiqued, especially when they describe separate ideas. Or is one only allowed to respond to the last sentence?

[mathjak107 is a kinder soul than I and responded to the point. In this forum, both glides and reverse glides for post-retirement have been discussed in great detail, dismissing the "dumbest thing in the world" to Strick]

Though bananas are tasty, banana-flavored food is often not tasty at all.

Well, I guess Strick can chime in to defend him/herself, but I believe you are still missing the point of his/her post.  The entire point of the post is contained in the last sentence.  He's "seen the light," and is now saying he no longer subscribes to the ideas detailed in the first paragraph.

[JoJoP]

I'm going to chime in about HELOC's.  We had 2 Helocs, and both were frozen. We got them both about 10 years ago, when the banks were lending to anything that fogged a mirror.   Both were for 150K CL's, and attached to paid off properties at less than 50% loan to value, and  used to buy investment property.

They were frozen for different reasons.  One was frozen after our banker suggested a credit line increase on it, and the Underwriting Department reviewed our income (self employed, sporadic), and froze it.  We had just paid it down by $40K, with the thought that we'd pull that 40K back out later in the year, when we needed it, but in the meantime it would effectively reduce our mortgage debt until then.  We did eventually get the $40K back out, but had to jump through numerous hoops to do so. 

The second Heloc was frozen when Washington Mutual was bought out by Chase.   We'd just gone through the episode above, and had no desire to use those helocs as the financial tools we had hoped for.  I still have it... it 's a 30 year pay off, at an under par interest rate.  It's still one of my lowest interest loans. 

All in all, though, as a means to acquire more rental properties, it was hugely successful, and I'd do it again if I had confidence that a bank would loan to us.  I think my point is that our financial picture, with self employment, no pensions, etc, hugely benefited from the access to easy money when it was available.  For us, it was a moment to be seized.    I think that opportunities for successful retirement are dotted along everybody's path.  Helocs at a cheap rate, the chance to buy a property at a discount, 401K matching, a wildly successful stock investment, a rental, an inheritance... these chances come in all shapes and sizes.   Don't squander them, no matter what form they take.  I'm in my mid-50's, so FI without RE,  with a 'stache in the mid 7 figures and growing.  I read MMM because I always learn something here, and, because like the chart shows,  hardly anyone of any age, including a big swath of my friends,   has enough money to last them. 

[TFR]

Long time reader of the forums but commenting for first time.  This thread has seen of the most exhaustive discussions of the 4% rule, it was very interesting to read different perspectives.  While the fees issue is clear, there is a fundamental difference between the earlier Trinity study and recent times.  The supreme over-valuation of stocks in 1999-2000 followed by yield compression in bonds didn't have parallels even among the worst case 1996 retiree cohort, who at least had nice interest coming in from bonds at that time.  This case needs to be studied in depth, though it's only been 17 years of retirement for the 2000 retiree. 

On a broader level, using convervative tweaks to arrive at the 'worse case' failure edge in SWR using historical data, I did a small study as described here: http://tenfactorialrocks.com/hacking-the-retirement-calculators/

I find that 3.27% to be the real safe withdrawal rate from this exercise, even considering current historically low bond yields. I would appreciate your views on it.   

[mathjak107]

the 2000 retiree is still doing okay. not great but about on par with the group that retired in 1929 . michael kitce's took an in depth look at the 2000 and 2008 retiree .

this is the summary

EXECUTIVE SUMMARY

The 4% rule has been much maligned lately, as recent market woes of the past 15 years – from the tech crash of 2000 to the global financial crisis of 2008 – have pressured both market returns and the portfolios of retirees.

Yet a deeper look reveals that if a 2008 or even a 2000 retiree had been following the 4% rule since retirement, their portfolios would be no worse off than any of the other “terrible” historical market scenarios that created the 4% rule from retirement years like 1929, 1937, and 1966. To some extent, the portfolio of the modern retiree is buoyed by the (only) modest inflation that has been occurring in recent years, yet even after adjusting for inflation, today’s retirees are not doing any materially worse than other historical bad-market scenarios where the 4% rule worked.

Ultimately, this doesn’t necessarily mean that the coming years won’t turn out to be even worse or that the 4% rule is “sacred”, but it does emphasize just how bad the historical market returns were that created it and just how conservative the 4% rule actually is, and that recent market events like the financial crisis are not an example of the failings of the 4% rule but how robustly it succeeds!


https://www.kitces.com/blog/how-has-the-4-rule-held-up-since-the-tech-bubble-and-the-2008-financial-crisis/

[Monkey Uncle]

the 2000 retiree is still doing okay. not great but about on par with the group that retired in 1929 . michael kitce's took an in depth look at the 2000 and 2008 retiree .

this is the summary

EXECUTIVE SUMMARY

The 4% rule has been much maligned lately, as recent market woes of the past 15 years – from the tech crash of 2000 to the global financial crisis of 2008 – have pressured both market returns and the portfolios of retirees.

Yet a deeper look reveals that if a 2008 or even a 2000 retiree had been following the 4% rule since retirement, their portfolios would be no worse off than any of the other “terrible” historical market scenarios that created the 4% rule from retirement years like 1929, 1937, and 1966. To some extent, the portfolio of the modern retiree is buoyed by the (only) modest inflation that has been occurring in recent years, yet even after adjusting for inflation, today’s retirees are not doing any materially worse than other historical bad-market scenarios where the 4% rule worked.

Ultimately, this doesn’t necessarily mean that the coming years won’t turn out to be even worse or that the 4% rule is “sacred”, but it does emphasize just how bad the historical market returns were that created it and just how conservative the 4% rule actually is, and that recent market events like the financial crisis are not an example of the failings of the 4% rule but how robustly it succeeds!


https://www.kitces.com/blog/how-has-the-4-rule-held-up-since-the-tech-bubble-and-the-2008-financial-crisis/

I'm generally a believer in the adequacy of the 4% rule, but I have to point out that if you go by the default cFiresim inputs, 1966 failed pretty badly (as did all the other years between 1965 and 1969).  So saying that 2000 and 2008 were no worse than the worst years of the past is not necessarily comforting.  What would be more comforting is saying that the late '60s, as well as 2000 and 2008, were anomalously bad, and as long as things don't get that bad again near the beginning of your retirement, you'll probably be o.k.

[mathjak107]

a 95% success rate  counts on 1965/1966 failing along with a few others .

but what you have to remember is conventional planning usually has us planning until 90 or 95 .

statistically that is  low odds any of us will see those ages so in reality if playing statistics we really are running a much higher success rate portfolio wise since odds are we are planning for more years than we will likely need .

even a 95% success rate when combined with life expectancy success have the 4% safe withdrawal rate up close to 100% even though the portfolio runs 95%  . we have not even fiigured in human spending patterns in our older years changing . much of what we do and buy earlier in life we stop later on and that pays for increases in what we do continue on with . so study after study shows comes the later years we need a lot less inflation adjusting than calculating an adjustment yearly .

we have not needed a raise  yet going in to our 3rd year in retirement .  been the same budget the last 2 years . this year two big expenses will drop as i go on medicare and my supplement picks up the 480 a year gym membership too . that is a savings of thousands a year from what i pay now in health insurance on my own .

so it is likely we may get to go 4 years with no inflation adjustment so far .

we don't use a fixed 4% anyway as a spending guide . our withdrawal system is dynamic and changes yearly  . we just have found even though we could spend more we ended up not .

[Monkey Uncle]

Yes, I agree with all of that, which is why I said that I'm basically a believer in the adequacy of the 4% rule.  What I was taking issue with is Kitces' assurance that 2000 and 2008 retirees will be o.k. because those years were no worse than 1966. 

[Retire-Canada]

even a 95% success rate when combined with life expectancy success have the 4% safe withdrawal rate up close to 100% even though the portfolio runs 95%  .

Charts from Maizeman. Red is the "Go Broke" zone for a 30yr old FIREr. If you retire later in life at a 4%WR the red zone gets hard to see. We should be worrying about dying not going broke. ;)

[moof]

*snip*
This is very encouraging given that it assumes FIRE at 30 which is on the early end of the spectrum for many Mustachians and does not include Social Security.

Yep, I'm looking at 22 years from planned date to age 70 where my SSA benefits add up to $54k, which is a little more than my planned withdrawals.  5% withdrawal looks quite safe to me, for my situation.  I have the added buffer of knowing my kid should stop being a major ongoing cost less than halfway through my early retirement period.

[mathjak107]

even a 95% success rate when combined with life expectancy success have the 4% safe withdrawal rate up close to 100% even though the portfolio runs 95%  .

Charts from Maizeman. Red is the "Go Broke" zone for a 30yr old FIREr. If you retire later in life at a 4%WR the red zone gets hard to see. We should be worrying about dying not going broke. ;)

This is very encouraging given that it assumes FIRE at 30 which is on the early end of the spectrum for many Mustachians and does not include Social Security.

exactly what life expectancy are those charts supposed to be illustrating ?  life expectancy can vary greatly based on what is being looked at . if it is from birth it is the most skewed . 

as michael kitces points out :

"“life expectancy” can be a somewhat misleading term. Many people hear the term and think of it as a measure of how long they can “expect to live”. In reality, though, life expectancy is a measure of the average time a person within some particular population is expected to live. While the average is meaningful in many respects, it may not always provide the best measure for setting expectations about the actual age someone is likely to reach. Because mortality rates aren’t constant across a lifespan and the distribution of ages at death are heavily skewed (i.e., more people die old than young), commonly cited life expectancy measures—particularly life expectancy at birth, which is most often cited in the media—may result in misleading expectations.

For instance, a child born in 2014 has a life expectancy (average age at death) of 79. However, the median age of death for the same child is 83, and the modal (most common) age at death is 89! Given the shape of the distribution of ages at death (negatively skewed), it’s simply a mathematical fact that the mean is going to be lower than the median or the mode."

[matchewed]

even a 95% success rate when combined with life expectancy success have the 4% safe withdrawal rate up close to 100% even though the portfolio runs 95%  .

Charts from Maizeman. Red is the "Go Broke" zone for a 30yr old FIREr. If you retire later in life at a 4%WR the red zone gets hard to see. We should be worrying about dying not going broke. ;)

This is very encouraging given that it assumes FIRE at 30 which is on the early end of the spectrum for many Mustachians and does not include Social Security.

exactly what life expectancy are those charts supposed to be illustrating ?  life expectancy can vary greatly based on what is being looked at . if it is from birth it is the most skewed . 

as michael kitces points out :

"“life expectancy” can be a somewhat misleading term. Many people hear the term and think of it as a measure of how long they can “expect to live”. In reality, though, life expectancy is a measure of the average time a person within some particular population is expected to live. While the average is meaningful in many respects, it may not always provide the best measure for setting expectations about the actual age someone is likely to reach. Because mortality rates aren’t constant across a lifespan and the distribution of ages at death are heavily skewed (i.e., more people die old than young), commonly cited life expectancy measures—particularly life expectancy at birth, which is most often cited in the media—may result in misleading expectations.

For instance, a child born in 2014 has a life expectancy (average age at death) of 79. However, the median age of death for the same child is 83, and the modal (most common) age at death is 89! Given the shape of the distribution of ages at death (negatively skewed), it’s simply a mathematical fact that the mean is going to be lower than the median or the mode."

It's for a person aged thirty. Maizeman has previously explained these charts in another thread.

[maizefolk]

exactly what life expectancy are those charts supposed to be illustrating ?  life expectancy can vary greatly based on what is being looked at . if it is from birth it is the most skewed . 

as michael kitces points out :

"“life expectancy” can be a somewhat misleading term. Many people hear the term and think of it as a measure of how long they can “expect to live”. In reality, though, life expectancy is a measure of the average time a person within some particular population is expected to live. While the average is meaningful in many respects, it may not always provide the best measure for setting expectations about the actual age someone is likely to reach. Because mortality rates aren’t constant across a lifespan and the distribution of ages at death are heavily skewed (i.e., more people die old than young), commonly cited life expectancy measures—particularly life expectancy at birth, which is most often cited in the media—may result in misleading expectations.

For instance, a child born in 2014 has a life expectancy (average age at death) of 79. However, the median age of death for the same child is 83, and the modal (most common) age at death is 89! Given the shape of the distribution of ages at death (negatively skewed), it’s simply a mathematical fact that the mean is going to be lower than the median or the mode."

The proportion of the graph that is in the "dead" category is calculated using the cumulative probability of death from social security actuarial tables (https://www.ssa.gov/oact/STATS/table4c6.html) starting from the risk of death between your 30th and 31st birthday and moving on from there. The reason to do this in graphical form is that is does actually calculate the distribution of risks of death instead of trying to reduce it to a single number. So the simulations above do calculate the left hand skew of the distribution of death risk.

There certainly are lots of other factors that influence how long someone might expect the live. Starting with gender. The chart above is for a 30 year old man, for a 30 year old woman there would be a bit less gray space. Also all sorts of things like family histories of cancer and/or heart disease. So it's definitely a rough approximation. But I can at least assure you that the way the data is analyzed does take into account the shape of the distribution for death rates by age, not just a summary statistic like mean/median/mode.

[EscapeVelocity2020]

Haven't spent much time here, but I'm horrified to see 'hand drawn' looking graphics with no real explanation of assumptions become an acceptable way to give people life-changing information.  I suppose I'm from an earlier era where we had to read a 40 page study with a few dry graphics in order to be educated sufficiently on the subject matter and make my own judgments.  Just looking at the graphs, I have plenty of qualms with pretty pictures and can see how they can mislead.  Hopefully the pretty graphics encourage people to dig deeper in to what applies to their own future.  But it is a disservice that a lot of the time these graphs purport to shed some hard-won light on the user's expectations for the future.

[Classical_Liberal]

The graph simply represents data that is easily accessible to you on CfireSim and the SS actuarial tables.  Im sure the SS administration and the creators of Cfiresim have 1000's of pages of data from which those conclusions are being drawn.  I can't imagine that source data is off limits to you either. If you don't believe it look here and here.  Personally I love the graphic art!  But I also love a dose of healthy skepticism.

[farfromfire]

...
But it is a disservice that a lot of the time these graphs purport to shed some hard-won light on the user's expectations for the future.

How, exactly?
Seems like these graphs are shining a great deal of light, even if you do not like the somewhat whimsical graphical format. Whether the light they shed was "hard-won" is a question of how much effort was expended by their creator; while I have no direct knowledge of that, I am quite sure you do not either.

ETA:
maizeman, would you feel comfortable posting your source code? I would love to play around with the different parameters without bugging you,

[Retire-Canada]

Haven't spent much time here, but I'm horrified to see 'hand drawn' looking graphics with no real explanation of assumptions become an acceptable way to give people life-changing information.

First off the blame for lack of context is my fault. Maizeman posted these charts in another thread to help illustrate the topics discussed there. I can't lay my finger on the link and because there are so many "OH MY GOD The 4% is NOT GOING TO WORK!" threads on this forum it's a bit hard to search for. If anyone recalls the original thread's link please post it.

Second shake your head if you expect a deep explanation of the details behind everything that's shared on this site that could potentially have a dramatic effect on a reader's life. That's a ridiculous expectation of a forum like this.

Third what I would expect from a reader's reaction to viewing this chart is simply that they put into perspective the risks of going broke vs. dying when they think about FIRE planning. That would be the launching off point for examining their current assumptions and validating them. None of financial information or life expectancy information around FIRE is secret and cannot be looked at independently.

People casually talk about the 4% rule without mention of the Trinity Study or any of the assumptions behind it on this site. Often they take liberties with the assumptions and leave out key details. We don't freak out about that. Why? Because we expect that anyone who is going to make a major course change on their life will not do so based on a random post from a stranger on an internet forum. But, they may get inspired to look into a topic they wouldn't have know about previously and that may lead to something life altering, which is great.

[TomTX]

Haven't spent much time here, but I'm horrified to see 'hand drawn' looking graphics with no real explanation of assumptions become an acceptable way to give people life-changing information.  I suppose I'm from an earlier era where we had to read a 40 page study with a few dry graphics in order to be educated sufficiently on the subject matter and make my own judgments.  Just looking at the graphs, I have plenty of qualms with pretty pictures and can see how they can mislead.  Hopefully the pretty graphics encourage people to dig deeper in to what applies to their own future.  But it is a disservice that a lot of the time these graphs purport to shed some hard-won light on the user's expectations for the future.

I think it is a perfectly splendid set of graphs, and the style makes it very accessible.

You must also be turning up your nose at XKCD which has some extremely insightful pages.

[matchewed]

So your posts will contain all the necessary background research it took to reach whatever conclusion or point you're trying to illustrate? And of course will look far more professional, whatever that means.

Was there ever such an era?

[sol]

Maizeman posted these charts in another thread to help illustrate the topics discussed there. I can't lay my finger on the link and because there are so many "OH MY GOD The 4% is NOT GOING TO WORK!" threads on this forum it's a bit hard to search for. If anyone recalls the original thread's link please post it.

That discussion was in this thread, with maizeman making his seminal contribution on page 2.

[maizefolk]

Edit (4/23/17): I've continued to clean up the code and push out more of my "helper scripts." More recent and generally easier to read versions of the code posted below are now on github: https://github.com/maizeman/dead_broke/tree/master/DAB_graphs

Sure thing FarFromFire. I'm attaching three files, one with the python code for actually generating the graphs, and the other two are example csv files carrying the data on mortality rates for people (from SSA) and portfolios (calculated using the shiller data). Any other human mortality data should work if you change it into the same csv format as "Death_rate.csv." Same for the portfolio withdrawal rate data (I had to regenerate the portfolio mortality rates for each withdrawal rate tested). The attached portfolio mortality rate file (broke_rates.csv) is for a 4% withdrawal rate.

In addition to tweaking the input files there are two configurable variables at the start of the python script. The first is "fire_age" and should be the age at which portfolio withdrawals start. The second is "male" and is either true which means the mortality data from men in the death_rate.csv spreadsheet is used, or false in which case the mortality data from women in the death_rate.csv spreadsheet is used.

In response to a previous question from ARS, I also posted the portfolio mortality rates for 100% stock portfolios between 3.5% and 6.0% in a more standard excel sheet.

This code was definitely written as an exoskeloton, not a robot,* so if you run into bugs running it on your computer, let me know.

Also, just for fun: If there are concerns that the the graphs don't look like the ones one would find in a 40 page peer reviewed paper from back when there was a per figure charge for including color and this makes people less likely to take the results seriously, that's a relatively straightforward issue to address.



I feel like it loses something in the translation though.

My personal convention is that, the more conceptual the idea I'm trying to get across, the more informal the style of the graph (so people don't try to extract exact values from it). The more quantitative the actual analysis, the more formalized the style.

*Professional programmers tend to design code that's like a robot, knowing that it will have to go out into the world and interact with people on its own (without the programmers' help if something goes wrong). Folks like me are much more likely to treat our code like exoskeletons. It increases our strength and abilities, but we're still there in the middle of it all with the ability to troubleshoot errors the code throws in real time.

[maizefolk]

Edit: EV2020, I do agree with you that having detailed and auditable methods and clear pointers to the raw data used for a given analysis is essential for both research and when people are making major life decisions. I didn't realize how many people would be interested in this form of data visualization when I ran a couple of quick models in the thread sol linked to above. Writing actual papers complete with cited sources and clearly defined introductions/methods/results/discussion is something I do in my day job, and the thought of doing the same thing for what I consider one of my fun hobbies that lets me decompress at the end of the day fills me with a feeling of exhaustion.  However, I have always been open about the datasets I used, and I have tried to be clear about methodology and assumptions (see below). Posting the python code actually used to generate the figures is another step in the right direction. (I'm still cleaning up the code used to generate the portfolio life expectancy table itself but hope to post that soon too.)

Gathered from various places around the forum.

The graphs are based off of essentially two numbers (let's leave aside the light blue vs dark blue distinction).

The SSA provides a number that is essentially the "risk of death" for a person in a given year based off of their age and gender. By looking at the inflation adjusted returns of the different 100% stock portfolio histories by starting in each different month of Shiller's dataset of 1871-present stock market returns, I calculated a "risk of bankruptcy" in a given year, using the number of years since a person started taking withdrawals from their portfolios and the withdrawal rate they've been using.

With those two numbers, it's just a matter of calculating proportions. At the start 100% of people are alive, 30 years old, none of them are bankrupt. Over the next year X% go bankrupt*, and 0.1505% of both bankrupt and non-bankrupt people die. Then at the start of the second year, 99.8405% of people are alive, 31 years old, and none of them are bankrupt. The graph continues that on for the next 70 years (so at the end 0.57% people are alive and 101 years old and 0.05% of people are alive and broke).

Age at FIRE, gender (different life expectancy), and withdrawal rate used are all going to chance the shape of the graph. In principle it'd also be possible to generate these for different investment mixes but *shrug* gotta draw the line somewhere.

*Actually with a 4% withdrawal rate and 100% stocks, no one manages to go bankrupt in year 1, but hopefully you see my point. With a 4% withdrawal rate the first year with a non-zero bankruptcy probability is year 17. Of the 1,546 start months with enough history to calculate stock returns out 17 years, one drops below zero during that year, giving a bankruptcy risk rate of 0.065% for 17 years into FIRE with a 4% withdrawal rate.

Things to know that mean my projections may not line up with the standard ones you might get out of a website like cfiresim: 1) I use monthly data on stock returns (shiller data) to calculate a lot more total scenarios. 2) that means I also calculate withdrawals on a monthly basis, not one lump sum per year 3) to calculate failure rates I use portfolio life expectancy, which I think provides a more accurate estimate of failure rates over extremely long retirements than looking at the number of failures out of all the time intervals that are as long as your estimate retirement in historical data (traditional Trinity approach).



My approach is the green line, the traditional approach is the blue line. Intuitively, failure rates shouldn't decrease as we go to longer retirement lengths, but the problem is that really bad years (like the mid 60s) start dropping out of your dataset once your retirement window gets long enough.

[farfromfire]

Sure thing FarFromFire. I'm attaching three files, one with the python code for actually generating the graphs, and the other two are example csv files carrying the data on mortality rates for people (from SSA) and portfolios (calculated using the shiller data). Any other human mortality data should work if you change it into the same csv format as "Death_rate.csv." Same for the portfolio withdrawal rate data (I had to regenerate the portfolio mortality rates for each withdrawal rate tested). The attached portfolio mortality rate file (broke_rates.csv) is for a 4% withdrawal rate.

In addition to tweaking the input files there are two configurable variables at the start of the python script. The first is "fire_age" and should be the age at which portfolio withdrawals start. The second is "male" and is either true which means the mortality data from men in the death_rate.csv spreadsheet is used, or false in which case the mortality data from women in the death_rate.csv spreadsheet is used.

Thanks! I'll be playing with it soon enough.

Also, just for fun: If there are concerns that the the graphs don't look like the ones one would find in a 40 page peer reviewed paper from back when there was a per figure charge for including color and this makes people less likely to take the results seriously, that's a relatively straightforward issue to address.

Oh sweet summer child:

The following are the basic charges for color:

    $1,045.00 US for printing charges (four or fewer color pages)
    $2,090.00 US (four to eight pages)
    $3,135.00 US (nine to twelve pages)
    + $62.50 US per color image (i.e., if you have two color figures the charge would be $125 US, + $1,045)

Online color is free, but they wouldn't accept figures in which color was necessary for the sake of print-only readers. IDK about other organizations/journals.

[maizefolk]

Yup, I think publication fees are evolving at different rates in different fields. I collaborate with a couple of folks in CS and they took the very idea of publication charges as a sign I wanted to publish in some sort of predatory fly by night journal. In my own field, the trend seems to be towards lumping per figure and per page charges into the overall publication fee. But my very first job in the lab I landed in from grad school was recreating a color bar chart in shades and patterns of gray, because, in the word of my PI at the time "either we can pay for a color figure, or I can buy another student a laptop."

All authors are assessed the following fees:

Regular research articles: $1,700 per article, with no additional fees for color figures or SI.
PNAS Plus articles: $2,300 per article, with no additional fees for color figures or SI.
Open access: Authors of research articles may pay a surcharge of $1,450 to make their paper freely available through PNAS open access option. If your institution has a site license, the open access surcharge is $1,100. All articles are free online after 6 months.

[farfromfire]

"either we can pay for a color figure, or I can by another student a laptop."

Lol, that's a great line.

Good on PNAS for moving to the 21st century and waiving color fees. OA fees don't affect me personally but hopefully that too will be gone one day.

[EscapeVelocity2020]

even a 95% success rate when combined with life expectancy success have the 4% safe withdrawal rate up close to 100% even though the portfolio runs 95%  .

Charts from Maizeman. Red is the "Go Broke" zone for a 30yr old FIREr. If you retire later in life at a 4%WR the red zone gets hard to see. We should be worrying about dying not going broke. ;)

This is very encouraging given that it assumes FIRE at 30 which is on the early end of the spectrum for many Mustachians and does not include Social Security.

Sorry for the freak out, but comments like this (and surely others that didn't comment) worry me that something as complicated and nuanced is being taken as 'there is a simple answer'. 

By 'hard-won', I was referring to the original Trinity study as well as the pile of follow on work from Pfau, Kitces, and many others.

Thanks for the measured response Maizeman, I wasn't meaning to call you out and I'm sure more work went on behind the scenes that might defeat the purpose of making the 4% rule easy to visualize.

[brooklynguy]

I think that someday the research (and computer simulations) will catch up with reality to verify that the 4% SWR with a variable withdrawal plan has been the right answer all along.

Kitces' site just published an outstanding article on "dynamic programming" (a/k/a "dynamic optimization") in the context of retirement planning, representing the convergence of financial-planning-practioner-oriented research and economist-oriented research as well as, in my view, an omen that the prediction expressed by Nords in the (nearly two-year-old) quote above (or, at the very least, the prediction that research (and computer simulations) will catch up to reality, whatever that reality may in fact be) is coming closer to fulfillment.

https://www.kitces.com/blog/dynamic-programing-irlam-tomlinson-methods-for-financial-planning-optimization/

[Jamese20]

right so i have read most the threads now on this 4%

I must say the "stop worrying about the 4%" aspect of this thread has really been counterproductive :) there is so many for and against points that a newer guy sorting his finances out over the next 12 months to then go all into preparing for FIRE just cant nail down a point in time when I can pull the trigger!

just merely looking at MMM's post i like the idea of 5% personally but there is so many of you well educated folks putting the crappers up me just on 4% that I feel like I am taking a huge risk!! :)

i can understand it because nobody knows how stocks are going to do in reality do they - so how do we plan on this? i would agree they always end up going up in the long haul but that still doesn't really mean it will continue.

at the moment i am just taking buffets advice and will just keep investing in businesses knowing they will keep growing and making money and i should do fine.

In summary, its almost unhelpful when there are so many conflicting viewpoints! :)

[matchewed]

right so i have read most the threads now on this 4%

I must say the "stop worrying about the 4%" aspect of this thread has really been counterproductive :) there is so many for and against points that a newer guy sorting his finances out over the next 12 months to then go all into preparing for FIRE just cant nail down a point in time when I can pull the trigger!

just merely looking at MMM's post i like the idea of 5% personally but there is so many of you well educated folks putting the crappers up me just on 4% that I feel like I am taking a huge risk!! :)

i can understand it because nobody knows how stocks are going to do in reality do they - so how do we plan on this? i would agree they always end up going up in the long haul but that still doesn't really mean it will continue.

at the moment i am just taking buffets advice and will just keep investing in businesses knowing they will keep growing and making money and i should do fine.

In summary, its almost unhelpful when there are so many conflicting viewpoints! :)

... it's like you've never even heard of the internet before. ;)

In all seriousness though, and this will go for any subject, it is up to you to keep educating yourself and use that knowledge to revisit this subject and come to your own conclusion. The conflicting viewpoints are actually rather regardless of you coming to a decision on a SWR.

[maizefolk]

Thanks for posting BG. We used to use dynamic programming (things like Needleman-Wunsch) for sequence alignments way back in the day. Things like identifying an ideal solution by starting with the endpoint and working backwards made sense. At that time the biggest point was that these actually reduced runtime complexity significantly compared to more naive approaches.

I'm still struggling to wrap my head around how you'd utilize a similar model when dealing with uncertain end points (unknown age at death, even though you'd have a probability distribution for it). Maybe that's where the stochastic dynamic programming models come in, which I'm not yet familiar with. Definitely calls for further reading.

This point was particularly well taken though:

While dynamic programming’s biggest strength is the ability to utilize some elegant models and complex math in order to optimize both asset allocation and distributions, arguably, this could also be one of dynamic programming’s biggest weakness. The insights of dynamic programming are only valuable so long as they reflect reality. If a utility function doesn’t actually capture a retiree’s utility or distributions aren’t actually as flexible as assumed, then the optimization under dynamic programming may not actually be optimizing utility and might even be suggesting actions that would decrease satisfaction in retirement.

The difficulty of building accurate utility functions that reflect the choices individuals actually make in the real world and the tendency of economists to eventually throw up their hands and say "okay we're going to model this as if people made logistically consistent decisions" is what finally got me to give up on economics back in school.

[Jamese20]

thanks for the notes virtus,

i guess it depends on how well you know the stock market and how well you think it will do and your confidence level in that.

i think during my saving decade i will just lump it all in my LS100 account and see what returns it did as a starting point

[Telecaster]

thanks for the notes virtus,

i guess it depends on how well you know the stock market and how well you think it will do and your confidence level in that.

It isn't even that complicated.  It really comes down to a question no one knows the answer to:  Will the investing future be worse than the past?   

History says 5% should be fine.  No one wants to recommend you take 5% because it in fact is a little riskier.   I don't recommend people drink beer.  I still do though.  It is a calculated risk.  The upside is a higher lifestyle.   The downside is you could run out of money at some point.  But as has been pointed out lots of times, it is possible to have a Plan B.  Some income on the side, downgrading your lifestyle, etc. 

[Eric]

thanks for the notes virtus,

i guess it depends on how well you know the stock market and how well you think it will do and your confidence level in that.

It isn't even that complicated.  It really comes down to a question no one knows the answer to:  Will the investing future be worse than the past?   

History says 5% should be fine.  No one wants to recommend you take 5% because it in fact is a little riskier.   I don't recommend people drink beer.  I still do though.  It is a calculated risk.  The upside is a higher lifestyle.   The downside is you could run out of money at some point.  But as has been pointed out lots of times, it is possible to have a Plan B.  Some income on the side, downgrading your lifestyle, etc.

Damn dude.  How much beer do you drink?

[dandarc]

thanks for the notes virtus,

i guess it depends on how well you know the stock market and how well you think it will do and your confidence level in that.

It isn't even that complicated.  It really comes down to a question no one knows the answer to:  Will the investing future be worse than the past?   

History says 5% should be fine.  No one wants to recommend you take 5% because it in fact is a little riskier.   I don't recommend people drink beer.  I still do though.  It is a calculated risk.  The upside is a higher lifestyle.   The downside is you could run out of money at some point.  But as has been pointed out lots of times, it is possible to have a Plan B.  Some income on the side, downgrading your lifestyle, etc.

Damn dude.  How much beer do you drink?

[cerat0n1a]

This thread and the other one that Maizeman provided the graphs for have been fantastic for clarifying my thinking. It really brings home that in my situation, I should care more about early death, ill health, divorce, effects of global warming, war, politics than about running out of money with a 4% withdrawal rate.

[dandarc]

This thread and the other one that Maizeman provided the graphs for have been fantastic for clarifying my thinking. It really brings home that in my situation, I should care more about early death, ill health, divorce, effects of global warming, war, politics than about running out of money with a 4% withdrawal rate.

I'd probably stick to caring most about early death, ill health, and divorce, as those areas you have the most control over.

[clumlee]

So with a 70/30 split what's the breakdown of the 70 between domestic/international and cap sizes?

[Monkey Uncle]

So with a 70/30 split what's the breakdown of the 70 between domestic/international and cap sizes?

That could fill a whole 'nother thread without reaching any kind of consensus.  But most folks will say that you should have at least some diversification across countries and capitalizations.

[EscapeVelocity2020]

This thread and the other one that Maizeman provided the graphs for have been fantastic for clarifying my thinking. It really brings home that in my situation, I should care more about early death, ill health, divorce, effects of global warming, war, politics than about running out of money with a 4% withdrawal rate.

I'd probably stick to caring most about early death, ill health, and divorce, as those areas you have the most control over.

And again, I'll just point out that the graphs are a gross oversimplification of reality.  The most important assumption is that history repeats itself or remains stagnant - lifetimes don't become longer, the worst 30-year rolling period for investors is behind us, etc.  At the very least, I encourage folks on this thread to visit FIREcalc, cFIREsim, and maybe i-ORP and Retirement Research.  Hopefully none of these links are spammy, but the internet seems to be turning in to a cash machine as opposed to free exchange of ideas it used to be...  alas, how can your fault a site owner that gets tens or hundreds of thousands of hits a day for monetizing it.  Just sucks that these resources are getting harder for the next generations to enjoy freely.

[cerat0n1a]

And again, I'll just point out that the graphs are a gross oversimplification of reality. 

Indeed, there can be no guarantees in future investment returns, or anything else, and the approach of using past years' data can have only limited predictive ability. Actuarial charts give death rates for the population, not for me. I expect to live longer than average, to be healthier than average and to avoid divorce, and I have some ability to influence those things, but the statistics suggest those are all more likely than running out of money.

At least on the finance side, there is also upside risk. My investments may do better than the worse case. The UK may continue to make good on its promise of a state pension in 20 years time for me and my wife. We may inherit money. Either of us might work for money at some point.

When looking several decades ahead, I do believe that technology, demographics (age profile of people in Europe for example) and climate change will have significant, difficult to predict impacts. Those are out of my control, but it's hard to see how an extra year of paid employment now helps with those.

[CanuckExpat]

So with a 70/30 split what's the breakdown of the 70 between domestic/international and cap sizes?

Complicated and controversial. No real way to tell you what will work best in the future without a crystal ball.

Take a look at this chart of the relative perfomance of various asset classes, taken from Callan Investment Returns Ranked by Asset Class 1997-2017, you can see the high performers vary a lot:


Then you have to take your personal situation into account regarding tax situation (if any advantage of domestic vs international), and also currency concerns.
When in doubt, going with roughly the market capitalization of each is cheap and easy and probably good enough. Most people tend to heavily overweight their home country. For US investors this might have worked out nicely in the last few years with the relative under-performance of international stocks, but whether that will carry on into the future, or is just an example of luck instead of home country bias will wait to be seen.

For what it's worth, I roughly split my stock allocation between US and "international" (I put international in quotes, since to other investors US is presumably international). I also tilt towards size and value, I have a slight over allocation to small cap and value stocks, as represented by whatever cheap indices are available to me.

[RapmasterD]

thanks for the notes virtus,

i guess it depends on how well you know the stock market and how well you think it will do and your confidence level in that.

It isn't even that complicated.  It really comes down to a question no one knows the answer to:  Will the investing future be worse than the past?   

History says 5% should be fine.  No one wants to recommend you take 5% because it in fact is a little riskier.   I don't recommend people drink beer.  I still do though.  It is a calculated risk.  The upside is a higher lifestyle.   The downside is you could run out of money at some point.  But as has been pointed out lots of times, it is possible to have a Plan B.  Some income on the side, downgrading your lifestyle, etc.

Damn dude.  How much beer do you drink?

This is a worthy question. Depending on sequence of returns in one's early years, she/he could drink a few bottles of Stone IPA with a grassfed boneless rib eye (good times) or a can or three of Keystone with an unadorned baked potato (not so good times). And then the money will last....oh it will last. Fear not. Drink. A lot.

[G. Thomas]

If you're interested in a cringe worthy Northwestern Mutual ad that makes some strong "what if" statements...

https://youtu.be/WBN_3sVWyUw

[PizzaSteve]

If you're interested in a cringe worthy Northwestern Mutual ad that makes some strong "what if" statements...

https://youtu.be/WBN_3sVWyUw

OMG those assumptions.

[nereo]

If you're interested in a cringe worthy Northwestern Mutual ad that makes some strong "what if" statements...

https://youtu.be/WBN_3sVWyUw

OMG those assumptions.

Basically he's talking about sequence of returns risk, and how the returns in the initial period are the ones that people ought to worry about.
BUt yes... the assumptions he was making was three or four consecutive years of 20% losses. 1930-32 is the only period that approximates this kind of incredible loss.  So much so that we've given it a name; the 'great depression'.

Kind of bizarre that he prefaces his argument by saying "things have changed in the last 22 years".  All told we've avoided the kind of multi-year bear markets he's talking about.
Then again, fear sells, and this is a bank presumably trying to sell financial services by making people too afraid to handle their own money.

[Classical_Liberal]

If you're interested in a cringe worthy Northwestern Mutual ad that makes some strong "what if" statements...

https://youtu.be/WBN_3sVWyUw

OMG those assumptions.

BUt yes... the assumptions he was making was three or four consecutive years of 20% losses. 1930-32 is the only period that approximates this kind of incredible loss.  So much so that we've given it a name; the 'great depression'.

Also realize that accompanying these astounding losses was a substantial amount of DEflation during those years.  So withdrawal amounts would have decreased to maintain standard of living.  $100 in January of 1930, was the same as $77 in December 1932.  His "scenario" was assuming massive year over year losses like that time period, but with inflationary spending increases.  Like taking a Monte Carlo calculator and using all the historical worse case economic inputs, even if they are reverse correlated, and creating a very low probability event to use as a baseline for predicting the future. 

[PizzaSteve]

Agreed.  And the irony is that after 3-4 years of 20% plus market losses, corporate defaults on debt, and a likely govt induced monetary system response, it is highly likely that the insurance company that sold you their 'guaranteed returns' structured product is belly up or using their wiggle room contract performance clauses to pay pennies on your dollar anyway.

[mathjak107]

the recovery from the great depression in dollar terms was less than 5 years .

the cpi fell 18% , dividends which hit 18% were not included and many very popular stocks like ibm were doing very well but not in any index yet .

in fact by the time the markets hit the same nominal levels many years later they were very very far ahead in dollar terms .

so just like comparing interest rates in times of higher inflation ,deflation has the reverse effect and makes recovery sooner

[Tyson]

Then again, fear sells, and this is a bank presumably trying to sell financial services by making people too afraid to handle their own money.

YES!  And not just with money, either.  If you ask people how safe are we (from violence), the answer is usually "not very" and often "Things are getting worse". 

Except the data shows that internal violence is down (ie, homcides), way, way down actually.  And death from wars (external violence) is way, way down too.  We live in just about the safest country in the world and the safest time in all of history. 

Yet people are still afraid.  It boggles my mind. 

And you even see it here on MMM - just hit one of the prepper threads.  For them, the fact that things are so good is proof positive that "its all coming down, man".  I just don't understand it. 

[nereo]

...
Yet people are still afraid.  It boggles my mind. 

And you even see it here on MMM - just hit one of the prepper threads.  For them, the fact that things are so good is proof positive that "its all coming down, man".  I just don't understand it.

I had a roommate who was a borderline prepper, but deeply entrenched in the prepper community.  There were times when it felt like they were actually rooting for the collapse of civilization so that they could feel good about all the resources they were stashing and fear they were spreading.

One of my guiding principles has been to never root against collective achievements. I don't want one country to fail just so that we're 'better' in comparison.
 ... rising tide floats all boats and whatnot.