4% SWR (or 3.x) has never been modeled with current interest rates

[bacchi]

Has anyone here not heard of VPW?   Variable percentage withdrawal.....from my perspective it alleviates many of the concerns being brought up here.

Of course we've heard of it. Here is a thread discussing it from a few years back:
https://forum.mrmoneymustache.com/investor-alley/variable-percentage-withdrawal-(vpw)/

cFIREsim also supports simulations with VPW. There's a pretty glaring downside of some very low income years.

VPW is still very safe with a floor.

McClung (Living Off Your Money) did research on high spend years. Portfolios can withstand a few emergencies and won't collapse. This is based on the fact that most 4% portfolios will grow to a very large amount.

[Boofinator]

VPW is still very safe with a floor.

McClung (Living Off Your Money) did research on high spend years. Portfolios can withstand a few emergencies and won't collapse. This is based on the fact that in most years 4% portfolios will grow have grown to a very large amount.

Slight (but in my opinion germane) correction.

[TomTX]

VPW is still very safe with a floor.

McClung (Living Off Your Money) did research on high spend years. Portfolios can withstand a few emergencies and won't collapse. This is based on the fact that in most years 4% portfolios will grow have grown to a very large amount.

Slight (but in my opinion germane) correction.

Follow that train of thought to its logical conclusion (can't trust anything based on the past) - and you're a Boglehead+, working and stashing (including food) until you die at your desk.

[TomTX]

Has anyone here not heard of VPW?   Variable percentage withdrawal.....from my perspective it alleviates many of the concerns being brought up here.

Of course we've heard of it. Here is a thread discussing it from a few years back:
https://forum.mrmoneymustache.com/investor-alley/variable-percentage-withdrawal-(vpw)/

cFIREsim also supports simulations with VPW. There's a pretty glaring downside of some very low income years.

Not if you put in a floor at 3.5%

Which is basically as safe as a flat 3.5% WR. Except you can spend 4% when the portfolio isn't below the inflation-adjusted starting value.

[RWD]

Has anyone here not heard of VPW?   Variable percentage withdrawal.....from my perspective it alleviates many of the concerns being brought up here.

Of course we've heard of it. Here is a thread discussing it from a few years back:
https://forum.mrmoneymustache.com/investor-alley/variable-percentage-withdrawal-(vpw)/

cFIREsim also supports simulations with VPW. There's a pretty glaring downside of some very low income years.

Not if you put in a floor at 3.5%

Which is basically as safe as a flat 3.5% WR. Except you can spend 4% when the portfolio isn't below the inflation-adjusted starting value.

Maybe I'm using this tool wrong but that doesn't look resilient at all... VPW with 3.5% floor compared with flat 3.5% SWR.

[TomTX]

Maybe I'm using this tool wrong but that doesn't look resilient at all... VPW with 3.5% floor compared with flat 3.5% SWR.

Yeah, something's wrong. Basically every start date is portfolio failure. I don't recall where I ran the simulations - it was awhile back. Also, be sure to cap spend at 4%.

[RWD]

Maybe I'm using this tool wrong but that doesn't look resilient at all... VPW with 3.5% floor compared with flat 3.5% SWR.

Yeah, something's wrong. Basically every start date is portfolio failure. I don't recall where I ran the simulations - it was awhile back. Also, be sure to cap spend at 4%.

Okay, I hadn't set the ceiling. Here it is with 3.5% floor to 4.0% ceiling. Looks like a slightly worse 3.5%, which sounds about right. With such a narrow window though it's not really going to offer much beyond a fixed withdrawal rate in practice.

[Boofinator]

VPW is still very safe with a floor.

McClung (Living Off Your Money) did research on high spend years. Portfolios can withstand a few emergencies and won't collapse. This is based on the fact that in most years 4% portfolios will grow have grown to a very large amount.

Slight (but in my opinion germane) correction.

Follow that train of thought to its logical conclusion (can't trust anything based on the past) - and you're a Boglehead+, working and stashing (including food) until you die at your desk.

Apologies for the late reply, the site kept crashing on me last week.

The problem with phrasing the 4% rule as something that will apply to the future with some level of certainty (and hence giving those new to the concept the impression that this is the case) is that it is counter to reality. Now, I'm not saying we should throw out the 4% rule, as it is one of the best things we have to predict a SWR. But what I'm saying is there is nothing even close to a guarantee that the 4% rule will hold in the future. Here's a few reasons:

1) Even in a completely random (stochastic) process (which the stock and bond markets most certainly are not), you need a robust data set to come to a conclusion with some certainty. The examples of gambler's fallacies are legion, so I don't want to get too deep into the woods here, plus I'm afraid any example I do come up with that would mimic past stock returns would be too complex to be of much good. That said, a researcher who analyzed a random binomial process of unknown success rate with 19 out of 20 successes (when you consider retirement periods as 5-year rolling periods), would not be able to statistically exclude a 20% failure rate with 95% confidence.

2) The 4% rule does not apply to individual portfolios, but to years' worth of portfolios. Basically, every person who retired using the 4% rule in the late 1960's (which up until that point would have had a 100% success rate) would have been looking for work two decades later when it became apparent that the 100% success rate for a 4% withdrawal strategy should be revised to a 95% success rate. Since by this time the portfolio is depleted and they've been out of work for 20 years, good luck going forward. There's nothing saying that even a 3.5-4% VPW might not suffer a similar fate in the future.

3) Most importantly, paradigms can change. Let's take a look at baseball homeruns. Up through 1919, there had never been a batter who hit 30 or more homeruns in a season (that's 49 seasons of MLB, if you're counting). Now, if I was a manager and was making strategic decisions, I would assume this trend would likely continue to a strong degree. But in fact, since 1920 there has been at least one player who has hit 30 or more homeruns every season (and these days there are dozens (54 currently)). Now, was I wrong in 1920 to base my strategy off of a lesser number of homeruns? No, there were probably very few people who would have seen the paradigm shift coming. Would I have been wrong if I had said (following the 1919 season) that most years would never have a 30 homerun player? Absolutely.... To relate this back to the stock market: if 30 years ago, someone said there was a possibility of Japan's economy only growing by 3.3% real over the next three decades (among a time of peace and prosperity), he/she would have been laughed out of town.

I don't want to sound like a Debbie downer here, since I believe the 4% rule is a good risk-based strategy to use as it's based off solid data. But to state that it will apply going forward the same as it has in the past is either a slip-of-tongue/exaggeration (which I assumed) or a potentially consequential misunderstanding of how the world works (probably not for the poster, but for someone reading the comment).

[nereo]

so... it's different this time?

[vand]

The 4% withdrawal rate starts to fail with more regularity when you extend withdrawal period.

This should not really be a surprise; in the Trinity study was designed to study a 30 year withdrawal period, and total capital depletion at the end of 30 years was not an impediment to success (indeed it could be argued this is exquistely judged scenario).

But if you are planning on your portfolio surviving for 40, 50 years or more that many FIREers are planning then to the portfolio's ability to safely sustain ongoing withdrawals beyond 30 years you better be prepared to modify your assumptions. 

Like principle on a mortgage payment over 30 years, the dropoff in principle on a FIRE portfolio is not a linear progression; there should be barely any dropoff in the early years while towards the end you are hacking chucks out of it.

[Boofinator]

More precisely: it might not be the same this time.

Doesn't mean it won't apply, just that there's nowhere near the evidence to say that it will with any sense of certainty (and there's ample evidence to suggest it won't apply on the same terms it currently rests (95% success rate of 4% rule over future 30-year periods)).

Granted there's some probability this won't change for generations to come, but I wouldn't bet my stash on it.

[secondcor521]

What would you suggest as a retirement strategy instead?

[Boofinator]

What would you suggest as a retirement strategy instead?

I said I like the 4% rule and the VPW applied to it is an even better strategy. All I wanted to convey with my original comment was saying that the 4% rule will apply with some certainty of success in the future is misleading (in my opinion). Apologies for being overly anal.

On a completely different topic: Is anyone either familiar with the research or has anyone done the math themselves that relates the data from cFIREsim (or the Trinity study, etc.) to some number of independent 30-year time periods? When I visit cFIREsim, it shows the years from 1871-2015, for something like 115 30-year periods; I assume for newer papers, we're getting close to 121 30-year periods, so let's use this figure for a prettier number. Now obviously, one can say there are at least 5 independent 30-year periods (for 150 years of total coverage), and no more than 121 independent 30-year periods. 5≤x≤121, where x represents the number of equivalent independent 30-year periods with similar statistical properties as true independent 30-year periods when using a rolling 30-year period. Is anyone familiar with a good value for x? I don't want to plow through a bunch of math if the answer exists out in the Mustachian hive mind. Thanks!

[markbike528CBX]

What would you suggest as a retirement strategy instead?

I said I like the 4% rule and the VPW applied to it is an even better strategy. All I wanted to convey with my original comment was saying that the 4% rule will apply with some certainty of success in the future is misleading (in my opinion). Apologies for being overly anal.

On a completely different topic: Is anyone either familiar with the research or has anyone done the math themselves that relates the data from cFIREsim (or the Trinity study, etc.) to some number of independent 30-year time periods? When I visit cFIREsim, it shows the years from 1871-2015, for something like 115 30-year periods; I assume for newer papers, we're getting close to 121 30-year periods, so let's use this figure for a prettier number. Now obviously, one can say there are at least 5 independent 30-year periods (for 150 years of total coverage), and no more than 121 independent 30-year periods. 5≤x≤121, where x represents the number of equivalent independent 30-year periods with similar statistical properties as true independent 30-year periods when using a rolling 30-year period. Is anyone familiar with a good value for x? I don't want to plow through a bunch of math if the answer exists out in the Mustachian hive mind. Thanks!

I believe @maizeman has brought up this issue and done the math on it.
But the stock market is likely to be correlated year to year in the future, so all the math is already built into the 4% rule.
It is one of the reasons Monte Carlo simulations don't simulate the market very well.

The problem with phrasing the 4% rule as something that will apply to the future with some level of certainty (and hence giving those new to the concept the impression that this is the case) is that it is counter to reality. Now, I'm not saying we should throw out the 4% rule, as it is one of the best things we have to predict a SWR. But what I'm saying is there is nothing even close to a guarantee that the 4% rule will hold in the future. Here's a few reasons: 

If the whole of the screwed up things that the Trinity Study and 4% rule does include to be good to 95%, what more certainty could you want?
There IS no 100% certainty.   To include your baseball analogy, I'd be pretty dang happy to be batting 0.95!    So the percentage might go down to 90% in the future, because.. this time is different.  So, that is still better certainty than I get out of most things in life. 

[CCCA]

What would you suggest as a retirement strategy instead?

I said I like the 4% rule and the VPW applied to it is an even better strategy. All I wanted to convey with my original comment was saying that the 4% rule will apply with some certainty of success in the future is misleading (in my opinion). Apologies for being overly anal.

On a completely different topic: Is anyone either familiar with the research or has anyone done the math themselves that relates the data from cFIREsim (or the Trinity study, etc.) to some number of independent 30-year time periods? When I visit cFIREsim, it shows the years from 1871-2015, for something like 115 30-year periods; I assume for newer papers, we're getting close to 121 30-year periods, so let's use this figure for a prettier number. Now obviously, one can say there are at least 5 independent 30-year periods (for 150 years of total coverage), and no more than 121 independent 30-year periods. 5≤x≤121, where x represents the number of equivalent independent 30-year periods with similar statistical properties as true independent 30-year periods when using a rolling 30-year period. Is anyone familiar with a good value for x? I don't want to plow through a bunch of math if the answer exists out in the Mustachian hive mind. Thanks!

I believe @maizeman has brought up this issue and done the math on it.
But the stock market is likely to be correlated year to year in the future, so all the math is already built into the 4% rule.
It is one of the reasons Monte Carlo simulations don't simulate the market very well.

The problem with phrasing the 4% rule as something that will apply to the future with some level of certainty (and hence giving those new to the concept the impression that this is the case) is that it is counter to reality. Now, I'm not saying we should throw out the 4% rule, as it is one of the best things we have to predict a SWR. But what I'm saying is there is nothing even close to a guarantee that the 4% rule will hold in the future. Here's a few reasons: 

If the whole of the screwed up things that the Trinity Study and 4% rule does include to be good to 95%, what more certainty could you want?
There IS no 100% certainty.   To include your baseball analogy, I'd be pretty dang happy to be batting 0.95!    So the percentage might go down to 90% in the future, because.. this time is different.  So, that is still better certainty than I get out of most things in life.

I agree there is no 100% certainty, except that a sub 3% withdrawal rate will perform better than 4%.  But that's not the main issue.  I think that the failures are most likely to occur as a sequence of returns risk in the early years.  IF you see that happening in your first decade of retirement, you should probably do something about it (cut back on spending or get some supplemental income), so that maybe your invariant plan would have failed, but your actual retirement does not.

[ChpBstrd]

We must humbly keep in mind that we are using the 20th century (and sometimes part of the 19th!) to predict the 21st century. Here are some key economically important things that seem to be different now.

1) Demographics. Western and Asian societies consist of an unprecedented percentage of elderly people. This means a higher tax or healthcare burden that is not offset by jobs or productivity gains. The percentage of the over-65 population in the US just surpassed where Japan was in the early 90s.

2) Corruption. The 20th century started out with relatively high levels of corruption but generations of muckraking newspaper journalists tackled the issue and the US became one of the least corrupt nations on earth. The impact of corruption on economic growth cannot be understated. Even today, the world’s poorest and richest economies lie on a continuum from most to least corrupt. Now that journalism is a vanishing profession, it is likely that developed economies could follow the economic performance of other nations that lack independent, professional journalism- like Russia.

3) Climate. What difference does an average 4-5 degrees F make? Well, the US economy is about to slowly lose all the infrastructure built and still being built in coastal cities like New Orleans, Tampa, and Miami, and cities like Houston and New York will soon face the existential challenges and insurmountable costs that now challenge New Orleans. Our trading partners will face similar costs, which will slow the growth of the global middle class as a market.

4) Financial Deregulation and Instability. The odds of portfolio success greatly improved after the Great Depression. I would argue this was due to the financial reforms that came out of the depression rather than the depression itself. Specifically, the Glass Steagal Act of 1933 marked the abrupt end of a hundred + year history of bank runs and contagious financial panics that had been translating the collapse of risky investments into a lack of lending for industrial purposes. Once investment banking was firewalled from commercial banking, the US enjoyed an unprecedented six decades without a liquidity crisis. Glass-Steagal was repealed in 1999 at the request of bank lobbyists and I think the results since then speak for themselves.

5) Debt. The US did not start the 20th century by going $23 trillion in debt. At some point, the majority of taxes paid or dollars created by the fed will go toward interest payments instead of economically valuable services such as education, infrastructure, salaries, or health. The most sensible way out of of the debt trap would be a currency devaluation/ debt monetization by fiat - that approach beats what the Greeks have had to endure. But that eventuality brings with it a lot of disruptive potential. Ask an Argentine.


Of course, the future could turn out in a way that mitigates all these issues, but each of these macro factors will in one way or another absorb trillions of dollars in economic production.

Yes, I’ve seen the wall of worry chart, and I acknowledge a lot of factors in the macroeconomy’s favor compared to the 20th century such as widespread literacy, workforce participation by women, and the promise of several new technologies such as construction by 3D printing, gene editing, AI, renewable energy tech, reusable rockets, and improved health informatics. The point is the 21st century will not be a repeat of the 20th, and we will only be able to calculate 21st century WR success rates on our death beds.

Retirement at any WR is a pure gamble, but one we must take.

[CCCA]

We must humbly keep in mind that we are using the 20th century (and sometimes part of the 19th!) to predict the 21st century. Here are some key economically important things that seem to be different now.

....

3) Climate. What difference does an average 4-5 degrees F make? Well, the US economy is about to slowly lose all the infrastructure built and still being built in coastal cities like New Orleans, Tampa, and Miami, and cities like Houston and New York will soon face the existential challenges and insurmountable costs that now challenge New Orleans. Our trading partners will face similar costs, which will slow the growth of the global middle class as a market.

...

I think the net effect of climate change is one of those things that is unclear.  While it will invariably have costs that we don't want to bear, lots of those costs could be counted as economic growth and contribute to GDP.  Just like every car accident and divorce can increase gpd (hospitals, car repair and lawyers, and extra housing all lead to higher GDP), climate mitigation and adaptation will require new energy technologies, greater energy use to cool buildings, costs to rebuild/relocate due to sea level rise, work to rebuild storm damaged areas, etc.  Not saying climate change is a good thing, there will definitely be many, especially low income folks, who are much worse off, but it's not clear (to me) that the stock market will be.

[Boofinator]

Just like every car accident and divorce can increase gpd (hospitals, car repair and lawyers, and extra housing all lead to higher GDP)....

Are you saying that there's literally nothing else that that money would have been spent on if it hadn't been for the car accident / divorce? Also keep in mind these kinds of incidents (moreso car accidents than divorces) take people away from actual productive employment, thereby potentially increasing inflation and degrading real increases in GDP.

[waltworks]

I think of it this way: I can spend X years of my life, which I cannot get more of, to earn and save money. I can then use that money to NOT work if I want.

If I just never stop working, my portfolio won't fail, but that's not the result I want.

If I quit today and withdraw 10% a year, that's almost certain to leave me destitute. Also undesirable.

So the key is to decide at what point I'm comfortable rolling the dice that my money will last long enough. Aiming for >95% success rates means you're almost certainly spending a significant chunk of your life working that you don't need to - a sort of front-loaded FIRE failure, if you will.

-W

[RWD]

Just like every car accident and divorce can increase gpd (hospitals, car repair and lawyers, and extra housing all lead to higher GDP)....

Are you saying that there's literally nothing else that that money would have been spent on if it hadn't been for the car accident / divorce? Also keep in mind these kinds of incidents (moreso car accidents than divorces) take people away from actual productive employment, thereby potentially increasing inflation and degrading real increases in GDP.

Yeah, this sounds suspiciously like the broken window fallacy...

[Bloop Bloop]

What would you suggest as a retirement strategy instead?

It depends on your work tolerance - each person can only speak for himself or herself. For me I like the security of knowing that my yearly spend will be covered wholly by dividends and rent without dipping into capital (plus my partner's income which I have not sought to allocate - I'm sure she can figure out something worthwhile for it). But I am happy to work a 20 year career  instead of the 10 year career I could have had if I was going to do a bare 4% rule retirement.

[waltworks]

Jesus. 10 years of your life is non-trivial!

I guess if you really like your job, maybe.

-W

[Bloop Bloop]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

[FIRE@50]

I feel like this thread could have been titled "6% SWR has never been modeled with current interest rates" Why be so insistent and being negative? No one can predict the future. Pick a model you are comfortable with and go with it.

[waltworks]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

Just for reference, if you're an adult who's been working for a decade or so (ie, in your 30s) you don't even have close to a 95% chance of surviving for 30 years, portfolio be damned. You're already at about a 15% risk of not being able to enjoy your retirement due to death by then.

Point being, life is shorter than you think and you can't get it back.

As an aside..."splurge"?!? Man, back in the day that would have earned you some major facepunching around here. Real MM'ers eat bark if they have to to get out of the cubicle a day sooner!

-W

[secondcor521]

As an aside..."splurge"?!? Man, back in the day that would have earned you some major facepunching around here. Real MM'ers eat bark if they have to to get out of the cubicle a day sooner!

-W

"Bark?  Why, we only dreamed of eating bark!  Would've been tiramisu to us!"

https://www.youtube.com/watch?v=ue7wM0QC5LE

[Bloop Bloop]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

Just for reference, if you're an adult who's been working for a decade or so (ie, in your 30s) you don't even have close to a 95% chance of surviving for 30 years, portfolio be damned. You're already at about a 15% risk of not being able to enjoy your retirement due to death by then.

Point being, life is shorter than you think and you can't get it back.

As an aside..."splurge"?!? Man, back in the day that would have earned you some major facepunching around here. Real MM'ers eat bark if they have to to get out of the cubicle a day sooner!

-W

Not sure about your figures. The 2019 Australian actuarial life tables state that the life expectancy for a 32 year old male is 51.3 years. I.e the 50 percentile survival rate is 51.3 years. It is therefore unlikely that the 15th percentile survival rate would be as low as 30 years.

I would like the ability to splurge, to not worry about recessions or having to rein in my spending, to be largely immune to economic and political shocks, to afford luxuries and frivolities if I wanted (95% of the time I don't, but I like having the option). I'm not saying that my attitude should be generalised. But it's what I want. I don't greatly enjoy working but on balance it's far from the worst thing in the world and I'm happy to go through the process till my early 40s.

[nereo]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

Just for reference, if you're an adult who's been working for a decade or so (ie, in your 30s) you don't even have close to a 95% chance of surviving for 30 years, portfolio be damned. You're already at about a 15% risk of not being able to enjoy your retirement due to death by then.

Point being, life is shorter than you think and you can't get it back.

As an aside..."splurge"?!? Man, back in the day that would have earned you some major facepunching around here. Real MM'ers eat bark if they have to to get out of the cubicle a day sooner!

-W

Not sure about your figures. The 2019 Australian actuarial life tables state that the life expectancy for a 32 year old male is 51.3 years. I.e the 50 percentile survival rate is 51.3 years. It is therefore unlikely that the 15th percentile survival rate would be as low as 30 years.

The point which was (correctly) being made is that you have a 32 year old has a >5% of dying before s/he turns 62 (i.e. a less than 95% chance of surviving).  Actuary tables are mathematically sobering things; this one (compiled from CDC data) suggests the probability of a 32 dying within 30 years is 13.4%  If you are 40 it's over 20%.

There's no additional 'magic' from living off dividends.  The improved success is solely the result of using a lower WR.  IN a similar vein, rental properties carry their own risk that's far harder to model. Rental markets can crash, neighborhoods can go to pot and entire cities can retract over decadal time periods.  And of course your individual properties can become damaged or destroyed in ways that aren't recoverable through insurance.

That's not to say that living off rental income or dividends isn't a bad strategy - both will likely serve you well.  But risk of portfolio failure isn't averted - the risks have just shifted.

[2Birds1Stone]

https://retirementresearcher.com/4-rule-work-todays-markets/

"This suggests that in a lower interest rate world, a 3% withdrawal rate reflects something closer to a chance of success than a 4% withdrawal rate historically provided over the broad range of historical market environments."

Happy OMY y'all

[Davnasty]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

This seems to be a consistent theme with those who argue for a more conservative withdrawal rate. It's not just a fear of failure at 4%, it's the increased chance of being able to spend more later in life. Those who argue for 4%+ withdrawal rates typically seem to really believe the idea that additional spending doesn't bring additional happiness. To them working longer is purely a way to increase the chance of success, but to the former group it's a way to increase the chance of success AND spend more. I think this is true for a lot of people but not everyone admits it like BloopBloop.

[markbike528CBX]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

Just for reference, if you're an adult who's been working for a decade or so (ie, in your 30s) you don't even have close to a 95% chance of surviving for 30 years, portfolio be damned. You're already at about a 15% risk of not being able to enjoy your retirement due to death by then.

Point being, life is shorter than you think and you can't get it back.

As an aside..."splurge"?!? Man, back in the day that would have earned you some major facepunching around here. Real MM'ers eat bark if they have to to get out of the cubicle a day sooner!

-W

Not sure about your figures. The 2019 Australian actuarial life tables state that the life expectancy for a 32 year old male is 51.3 years. I.e the 50 percentile survival rate is 51.3 years. It is therefore unlikely that the 15th percentile survival rate would be as low as 30 years.

The point which was (correctly) being made is that you have a 32 year old has a >5% of dying before s/he turns 62 (i.e. a less than 95% chance of surviving).  Actuary tables are mathematically sobering things; this one (compiled from CDC data) suggests the probability of a 32 dying within 30 years is 13.4%  If you are 40 it's over 20%.

There's no additional 'magic' from living off dividends.  The improved success is solely the result of using a lower WR.  IN a similar vein, rental properties carry their own risk that's far harder to model. Rental markets can crash, neighborhoods can go to pot and entire cities can retract over decadal time periods.  And of course your individual properties can become damaged or destroyed in ways that aren't recoverable through insurance.

That's not to say that living off rental income or dividends isn't a bad strategy - both will likely serve you well.  But risk of portfolio failure isn't averted - the risks have just shifted.

And for those who like a graph, http://engaging-data.com/will-money-last-retire-early/  has death probabilities for all you OMY'ers.

Apparently Australians are a bit longer lived, (32+51years =61% death rate for North Americans), but that only reduces the probability death at 32+30 years to about 12% or so.  Still greater than the 5% we're agonizing about here.

[matchewed]

https://retirementresearcher.com/4-rule-work-todays-markets/

"This suggests that in a lower interest rate world, a 3% withdrawal rate reflects something closer to a chance of success than a 4% withdrawal rate historically provided over the broad range of historical market environments."

Happy OMY y'all

This has been covered before. That discusses Monte Carlo vs historical. Monte Carlo simulations will randomly pick returns and returns are not random.

[Villanelle]

It's like every week someone is shocked to discover that the 4% rule isn't infallible.

I'm absolutely fascinated by this.
No WR is safe no matter what mathematical model you use.
You can't Boglehead your way to total financial security.

The models are a best guess at how to make decisions today to anticipate the financial needs of tomorrow and to assess the possible risk mitigating effects of different options.

It's all just probabilities and not dissimilar to the weather forecast. The calculations behind it are solid and the best we've got for deciding whether or not to bring an umbrella.
Though, that decision will depend far more on the individual's willingness to risk getting wet.

Is it wise to retire on exactly 25X your base expenses and  to spend every single cent of your budget every single year, plus inflation, while blindly ignoring what the markets and global economy are doing, while keeping no doors open for future earnings if necessary???
Lol, probably not.

What these models do is allow us to decide today what mitigating strategies we think best fit our particular risk tolerances for what might happen tomorrow.

For one person, ageism might be a huge factor in their industry, so banking on going back to their career is a bad mitigating strategy. They may choose to be more conservative, or they may choose to build skills that aren't subject to ageism, or both.

For another, they might have citizenship and family in an extremely low cost geographic region and can easily geo-arbitrage their way to a much lower spend rate.

For yet another, they might not have a lot of flexibility and may have a lot of high fixed costs, say, for a medically complex child whose specialists are only in HCOL areas of a country with expensive healthcare. Their best bet may be an extremely low WR and a large cash reserve.

For someone whose stache is primarily in index funds, AA based strategies might be best for mitigating risks like SORR. For someone with a rock solid defined benefit pension that covers their entire bare bones spend, a bond tent or cash reserves might be overkill.

For someone whose stache is primarily in real estate...well, I have no idea what they might do since I don't read much about primarily RE strategies. But you get the point.

All plans can fail.
The best thing you can do is try to understand your possible and probable failures, know your particular risk factors, and modulate accordingly.

The 4% rule is a starting point, not an end goal.
If you just got a perm and wear a lot of silk, grab an umbrella. Me? I don't mind getting rained on and I don't worry about the 4% rule. YMMV

Yes, it's like someone playing Blackjack, hitting on 12, and then being shocked that they busted because they were well within the guidelines.

That's why, for a while, there was so much kerfuffle when people were saying "SWR" meant Safe Withdraw Rate (rather than "Savings..."). 

It's a guideline to help with decision making.  And the remaining uncertainty is why people discuss the possibilities of going back to work or cutting expenses.  Those things inherently acknowledge the fact that 4% isn't some magical number that gives immunity from all financial woes.  if 4% were absolutely infallible, we wouldn't need to talk about or consider those things. 

[Bloop Bloop]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

Just for reference, if you're an adult who's been working for a decade or so (ie, in your 30s) you don't even have close to a 95% chance of surviving for 30 years, portfolio be damned. You're already at about a 15% risk of not being able to enjoy your retirement due to death by then.

Point being, life is shorter than you think and you can't get it back.

As an aside..."splurge"?!? Man, back in the day that would have earned you some major facepunching around here. Real MM'ers eat bark if they have to to get out of the cubicle a day sooner!

-W

Not sure about your figures. The 2019 Australian actuarial life tables state that the life expectancy for a 32 year old male is 51.3 years. I.e the 50 percentile survival rate is 51.3 years. It is therefore unlikely that the 15th percentile survival rate would be as low as 30 years.

The point which was (correctly) being made is that you have a 32 year old has a >5% of dying before s/he turns 62 (i.e. a less than 95% chance of surviving).  Actuary tables are mathematically sobering things; this one (compiled from CDC data) suggests the probability of a 32 dying within 30 years is 13.4%  If you are 40 it's over 20%.

There's no additional 'magic' from living off dividends.  The improved success is solely the result of using a lower WR.  IN a similar vein, rental properties carry their own risk that's far harder to model. Rental markets can crash, neighborhoods can go to pot and entire cities can retract over decadal time periods.  And of course your individual properties can become damaged or destroyed in ways that aren't recoverable through insurance.

That's not to say that living off rental income or dividends isn't a bad strategy - both will likely serve you well.  But risk of portfolio failure isn't averted - the risks have just shifted.

Thanks for providing the link. I couldn't find any equivalent for Australian actuarial figures except these life tables which I was citing previously:
https://cumsar.com.au/wp-content/uploads/multipliers_ready_reckoner_2019.pdf

In my link, a male aged 54 has a life expectancy of 30 years, so that should equate to a 50% chance of dying. In your link, the figure is 58%. In my link, a male aged 68 has a medan expectancy of 20 years. In your link, a male aged 68 has a 68% chance of dying within 20 years. So I think your demographic data is more death-prone than mine, maybe because Americans have a whole have more problems, particularly in their lower class and particularly since the GFC. But anyway, I accept what you say that there is a significant chance that a young person might not live to 65.

Even so, I prefer taking a conservative route - it's not like my working life is terrible or that I am not taking holidays and breaks and trying to enjoy myself while working.

[waltworks]

"Not terrible" is not how I want to live my life, though...

If you love your job (I love mine and will probably keep working ~15-20 hours a week forever) then you can just ignore the whole FIRE/4% SWR thing - you're already doing what you love. Lots of people are like that.

Otherwise, going for super conservative WRs just means you're wasting the time when you're young and healthy so you can be even richer when you're old. Seems like a poor tradeoff to me.

-W

[TomTX]

https://retirementresearcher.com/4-rule-work-todays-markets/

"This suggests that in a lower interest rate world, a 3% withdrawal rate reflects something closer to a chance of success than a 4% withdrawal rate historically provided over the broad range of historical market environments."

Happy OMY y'all

Pfau has made a career of sandbagging the 4% WR in his calculations then publishing it as gospel. It certainly got your clicks.

Load up with 1% professional managment fee? Yep. That's one way Pfau has caused a "failure" of the 4% SWR.
Handicap the Monte Carlo simulation with returns consistently 25% worse than actual historical returns? Yep. Did that too.
Plus other approaches I'm not recalling at the moment. Dig in, see how he did it this time.

Not worth our time.

[TomTX]

I agree there is no 100% certainty, except that a sub 3% withdrawal rate will perform better than 4%.

Depends what you mean by "perform". If you mean "maximize the likely number of years retired" - it certainly does not perform better.

[2Birds1Stone]

Not worth our time.

I guess you haven't read the ERN WR series.

[Telecaster]

Not worth our time.

I guess you haven't read the ERN WR series.

I have, and found it hugely informative (seriously).  ENR  assumes conditions that have never existed in the past will exist in the future. 

I find those assumptions to be completely implausible. 

[Boofinator]

I agree there is no 100% certainty, except that a sub 3% withdrawal rate will perform better than 4%.

Depends what you mean by "perform". If you mean "maximize the likely number of years retired" - it certainly does not perform better.

If we used that metric, a 5%, 6%, or even 7% rule would be the norm.

[Davnasty]

I agree there is no 100% certainty, except that a sub 3% withdrawal rate will perform better than 4%.

Depends what you mean by "perform". If you mean "maximize the likely number of years retired" - it certainly does not perform better.

If we used that metric, a 5%, 6%, or even 7% rule would be the norm.

Maybe. If we get into that line of reasoning we also need to start taking into account income and savings rate. If I can go from a 4% WR to 3% with one extra year of savings it might be a good choice whereas someone with a lower salary may require 5-10 more years of work to get the same result.

I've always taken the 4% rule to be the number that works the best for the most people. For an unusually high earner or unusually low spender, a lower withdrawal rate makes more sense and vice versa. If someone works for minimum wage a 7% withdrawal rate may actually be the right choice, not only because a safer rate would take many more years but because they could replace that salary more easily.

But ya, I think CCCA was strictly referring to success rate.

[Davnasty]

What would you suggest as a retirement strategy instead?

I said I like the 4% rule and the VPW applied to it is an even better strategy. All I wanted to convey with my original comment was saying that the 4% rule will apply with some certainty of success in the future is misleading (in my opinion). Apologies for being overly anal.

On a completely different topic: Is anyone either familiar with the research or has anyone done the math themselves that relates the data from cFIREsim (or the Trinity study, etc.) to some number of independent 30-year time periods? When I visit cFIREsim, it shows the years from 1871-2015, for something like 115 30-year periods; I assume for newer papers, we're getting close to 121 30-year periods, so let's use this figure for a prettier number. Now obviously, one can say there are at least 5 independent 30-year periods (for 150 years of total coverage), and no more than 121 independent 30-year periods. 5≤x≤121, where x represents the number of equivalent independent 30-year periods with similar statistical properties as true independent 30-year periods when using a rolling 30-year period. Is anyone familiar with a good value for x? I don't want to plow through a bunch of math if the answer exists out in the Mustachian hive mind. Thanks!

I believe @maizeman has brought up this issue and done the math on it.
But the stock market is likely to be correlated year to year in the future, so all the math is already built into the 4% rule.
It is one of the reasons Monte Carlo simulations don't simulate the market very well.

The problem with phrasing the 4% rule as something that will apply to the future with some level of certainty (and hence giving those new to the concept the impression that this is the case) is that it is counter to reality. Now, I'm not saying we should throw out the 4% rule, as it is one of the best things we have to predict a SWR. But what I'm saying is there is nothing even close to a guarantee that the 4% rule will hold in the future. Here's a few reasons: 

If the whole of the screwed up things that the Trinity Study and 4% rule does include to be good to 95%, what more certainty could you want?
There IS no 100% certainty.   To include your baseball analogy, I'd be pretty dang happy to be batting 0.95!    So the percentage might go down to 90% in the future, because.. this time is different.  So, that is still better certainty than I get out of most things in life.

I agree there is no 100% certainty, except that a sub 3% withdrawal rate will perform better than 4%.  But that's not the main issue.  I think that the failures are most likely to occur as a sequence of returns risk in the early years.  IF you see that happening in your first decade of retirement, you should probably do something about it (cut back on spending or get some supplemental income), so that maybe your invariant plan would have failed, but your actual retirement does not.

This is a good point that I often see overlooked. We discuss the possibility of running out 20+ years after retirement or the fact that a retirement >30 years has different rules, but in cases where the 4% rule is going to fail, one would almost certainly see the writing on the wall within the first 5-10 years. I believe it was in this thread someone mentioned running out after 20 years and how hard it would be to find work after that long out of the job market, but that's not something I worry about. My greatest worry would be a big drop in the first 5 years and needing to find work at that time. Personally, I would probably look for something part time or otherwise low paying and low stress. In that scenario even a small income would make a huge difference to success rate as it would reduce the need to sell and perhaps even buy a few more shares in a down market. Heck, for the most flexible of us it might make sense to use a 10% withdrawal rate knowing they will probably need to work again someday unless they get real lucky. That would be the best way to truly minimize working years.

I completely understand why anyone on the higher end of the income and/or annual spend spectrum would aim for a lower withdrawal rate. In fact it seems the divide in these arguments about SWR is often tied very closely to income and annual spend; that is, people who earn high incomes or spend more tend to argue for lower withdrawal rates and vice versa (based on comments posters have made elsewhere in the forum about their personal finances I see a loose trend, could be wrong on that)

[fattest_foot]

This seems to be a consistent theme with those who argue for a more conservative withdrawal rate. It's not just a fear of failure at 4%, it's the increased chance of being able to spend more later in life. Those who argue for 4%+ withdrawal rates typically seem to really believe the idea that additional spending doesn't bring additional happiness. To them working longer is purely a way to increase the chance of success, but to the former group it's a way to increase the chance of success AND spend more. I think this is true for a lot of people but not everyone admits it like BloopBloop.

An alternate way of looking at it is that each additional year you work is a 2-3% increased chance of death (depending on your age).

My math is probably wrong there, based on each year not having an equal chance at death (nor does using actuarial tables reflect an individual).

But regardless, I think the concept should be considered. Each additional year of work not only is a year of retirement lost, it's also a year closer to death.

[Davnasty]

This seems to be a consistent theme with those who argue for a more conservative withdrawal rate. It's not just a fear of failure at 4%, it's the increased chance of being able to spend more later in life. Those who argue for 4%+ withdrawal rates typically seem to really believe the idea that additional spending doesn't bring additional happiness. To them working longer is purely a way to increase the chance of success, but to the former group it's a way to increase the chance of success AND spend more. I think this is true for a lot of people but not everyone admits it like BloopBloop.

An alternate way of looking at it is that each additional year you work is a 2-3% increased chance of death (depending on your age).

My math is probably wrong there, based on each year not having an equal chance at death (nor does using actuarial tables reflect an individual).

But regardless, I think the concept should be considered. Each additional year of work not only is a year of retirement lost, it's also a year closer to death.

I agree on both accounts, the chance of death really helps to put the risk of running out of money into perspective but at the same time actuarial tables are probably not the best representation of those who aim to FIRE. Something tells me a subgroup of the population who make long term plans to retire early and have incomes above the median probably have a greater life expectancy than the general population.

[Boofinator]

I agree there is no 100% certainty, except that a sub 3% withdrawal rate will perform better than 4%.

Depends what you mean by "perform". If you mean "maximize the likely number of years retired" - it certainly does not perform better.

If we used that metric, a 5%, 6%, or even 7% rule would be the norm.

Maybe. If we get into that line of reasoning we also need to start taking into account income and savings rate. If I can go from a 4% WR to 3% with one extra year of savings it might be a good choice whereas someone with a lower salary may require 5-10 more years of work to get the same result.

I've always taken the 4% rule to be the number that works the best for the most people. For an unusually high earner or unusually low spender, a lower withdrawal rate makes more sense and vice versa. If someone works for minimum wage a 7% withdrawal rate may actually be the right choice, not only because a safer rate would take many more years but because they could replace that salary more easily.

But ya, I think CCCA was strictly referring to success rate.

Couldn't agree more. For a person making low income who hates their job, he/she should lean much higher than 4%. For someone with high income and a love of their job, SWAMI status and a very low withdrawal rate awaits. For those of us closer to the middle of the bell curve, 4% is a good balance of risk and reward.

[Metalcat]

This is a good point that I often see overlooked. We discuss the possibility of running out 20+ years after retirement or the fact that a retirement >30 years has different rules, but in cases where the 4% rule is going to fail, one would almost certainly see the writing on the wall within the first 5-10 years. I believe it was in this thread someone mentioned running out after 20 years and how hard it would be to find work after that long out of the job market, but that's not something I worry about. My greatest worry would be a big drop in the first 5 years and needing to find work at that time. Personally, I would probably look for something part time or otherwise low paying and low stress. In that scenario even a small income would make a huge difference to success rate as it would reduce the need to sell and perhaps even buy a few more shares in a down market. Heck, for the most flexible of us it might make sense to use a 10% withdrawal rate knowing they will probably need to work again someday unless they get real lucky. That would be the best way to truly minimize working years.

I completely understand why anyone on the higher end of the income and/or annual spend spectrum would aim for a lower withdrawal rate. In fact it seems the divide in these arguments about SWR is often tied very closely to income and annual spend; that is, people who earn high incomes or spend more tend to argue for lower withdrawal rates and vice versa (based on comments posters have made elsewhere in the forum about their personal finances I see a loose trend, could be wrong on that)

There are a lot of factors.
A lot of high earners have invested A LOT in order to get where they are, so there can be a lot of attachment to the level of work that they do and the circles of people it involves them with.

A lot of high earners know that it would be extremely difficult to get back to their previous income level, so the prospect of ever having to go back to work is very unattractive.

A lot of high earners have a much higher spend, which logically should mean that they have a lot of fat to cut in down times, but if they are well hedonically adapted, that prospect is just as unpalatable as going back to work for a major pay cut.

So, if they have put a ton of energy into getting to their professional level, enjoy the work (enough) and take a lot of pride from it, have entrenched high spending patterns, and know that once they let that income go, that it's pretty much gone forever...yeah, more years working would seem like a small price to pay compared to FIRE failure.

Comparably, someone who always hated their job and took little pride in it and may have no emotional connection to their professional community. If that same person has a low spend that could easily be covered/subsidized by casual work that they can easily hop in to, then minimal increase in risk of FIRE failure just isn't motivating.

The risk worth paying attention to and accounting for is whatever risk the person most wants to avoid.

[ChpBstrd]

Maybe. If we get into that line of reasoning we also need to start taking into account income and savings rate. If I can go from a 4% WR to 3% with one extra year of savings it might be a good choice whereas someone with a lower salary may require 5-10 more years of work to get the same result.

To go from a portfolio with 25x spending to one with 33x spending, one would have to save 7x spending from some combination of portfolio gains and salary. To figure out how long this would take, solve for X in the following equation (which produces a valid estimates only for short periods because it ignores inflation and compounding!):

(X years)Portfolio return + (X years)Savings rate = (7)Spending

Stated another way:

X years (Portfolio Returns as a factor of spending + Savings Rate as a factor of spending) = 7x
Or...
7/(PR+SR) = X

We can assume the 25x portfolio is expected to gain 1x per year, so at a savings rate of zero we gain our 7x in about 7 years.

7/(1+0) = 7

 That high-end expectation can be improved upon by saving something. At a 50% savings rate, the PR + SR would be 1.5x per year and 7x would be reached in 4.7 years.

7/(1+0.5)=4.7

At a 70% SR, one would get there in 4.1 years. At a 30% WR, it would be 5.4 years. YMMV but the average Mustachian can expect to spend 4-5 years of their lives to get from a 4% WR to 3%.

The question is, could you skip working those 4-5y now but agree to start working them sometime in the next 5-10y only if a SORR event happens, and skip the extra work if it doesn’t? As we’ve seen, if one’s portfolio was decimated by 7x spending and went from 25x to 18x, a 28% decline and a 5.6% WR, one could earn that 7x back with about 5 years labor (and still be young enough to do so). But the probability of having to do those 5y labor is something less than 100% so isn’t that better than accepting the 100% probability of having to work those same years to hit 33x?

Also note that if one’s 33x portfolio lost the same 28%, one would be down to 23.8x, so perhaps one would need to put in a year or two of labor to get up to 25x. After all, that’s the standard we’ve set for the 25x retiree in the same scenario.

Option 1: 25x
     Retire now
     Return to work for 5Y if portfolio drops 28% in next X years (risk <100%)
     Total: 5y(risk)

Option 2: 33x
     Work 5 more years (risk=100%)
     Return to work 1Y if portfolio drops 28% in next X years (risk <100%)
     Total: 5y(100%) + 1y(risk)

[PhilB]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

This seems to be a consistent theme with those who argue for a more conservative withdrawal rate. It's not just a fear of failure at 4%, it's the increased chance of being able to spend more later in life. Those who argue for 4%+ withdrawal rates typically seem to really believe the idea that additional spending doesn't bring additional happiness. To them working longer is purely a way to increase the chance of success, but to the former group it's a way to increase the chance of success AND spend more. I think this is true for a lot of people but not everyone admits it like BloopBloop.

I wonder how often the 3% vs 4% argument is just a failure of communication.  Mr 4% says I'm happy my $1M stash will throw off $40k and not worried about SOR, etc because really I'd be perfectly happy living on $30k so cutting back wouldn't be an issue.  Mr 3% says his $1M stash will only reliably throw off $30k as that's the amount I need to be happy and  I wouldn't be able to face cutting below it.

The only real difference between the two is that Mr 4% is giving himself permission to spend some of that probable extra fat now if markets look okay, Mr 3% is deferring that until some future date. 

[Metalcat]

My job is not the worst, in fact objectively I struggle to think of any way my job could be better, other than by paying more.

10 years is non-trivial but so is the knowledge for the rest of my life that I am financially very sound, and I am able to splurge pretty heavily.

This seems to be a consistent theme with those who argue for a more conservative withdrawal rate. It's not just a fear of failure at 4%, it's the increased chance of being able to spend more later in life. Those who argue for 4%+ withdrawal rates typically seem to really believe the idea that additional spending doesn't bring additional happiness. To them working longer is purely a way to increase the chance of success, but to the former group it's a way to increase the chance of success AND spend more. I think this is true for a lot of people but not everyone admits it like BloopBloop.

I wonder how often the 3% vs 4% argument is just a failure of communication.  Mr 4% says I'm happy my $1M stash will throw off $40k and not worried about SOR, etc because really I'd be perfectly happy living on $30k so cutting back wouldn't be an issue.  Mr 3% says his $1M stash will only reliably throw off $30k as that's the amount I need to be happy and  I wouldn't be able to face cutting below it.

The only real difference between the two is that Mr 4% is giving himself permission to spend some of that probable extra fat now if markets look okay, Mr 3% is deferring that until some future date.

This is what I come back to over and over and over again.

People can debate these little percentage differences all they want, but the variability in the outcome of the simulation is peanuts compared to the variability inherent in the inputs.

That's why I find it all so laughable. It's real math done with made up numbers. The outputs are nonsense beyond the vaguest of relevance to actually future outcomes.

If a "100% success" output from a simulator makes someone feel better, then that's cool, good for them, but since they'll never ever actually live according to the reality of their inputs, it's all gibberish and about as accurate as a BuzzFeed quiz about which Kardashian is my soulmate.

You can save to the exact dollar amount that will give you that exact 3.37875% WR that you calculated a million times, and that's cool, but no matter what, if you are inflexible in your spend, you are at substantially more risk of failure.

That's not reality though, no one is going to sit there with the global markets collapsing around them, look at their mangled 'stache, and say "nope nope nope, an online calculator on the internet that I used 30 years ago told me I can spend exactly $67K+estimated inflation, so goddammit I'm gonna spend that this year! Get packed kids! We're going on a cruise!"

If anything, these remarkably small impacts of WR reductions should reassure people that saving enormous sums more is almost irrelevant. Instead of freaking people out, it should really reassure people that over saving is nowhere near as powerful as common sense flexibility.

A very low WR is really great for one specific, highly probable scenario: dying with A LOT more money than you retired with.

Over saving is fine. I see no fundamental merit in saving only exactly what one anticipates needing. Pete has plenty more than he needs, I'll have plenty more than I need.

What's absurd is not feeling "safe" to start really living until some online simulation spits out "100%" and concluding that that actually has any bearing on predictability of the future.
That's just patently insane.

[chevy1956]

People can debate these little percentage differences all they want, but the variability in the outcome of the simulation is peanuts compared to the variability inherent in the inputs.

It all comes down to your estimated expenses doesn't it. In reality I see it as hitting a reasonable target (a 3%-5%) with a reasonable estimate of expenses and then some buffer in some format. The buffer could be the ability to decrease expenses or it could be additional savings. Then you figure out your drawdown plan for at least the first say 5 years and then you just give it a shot. It's just a rough estimate that hopefully has some flexibility in the plan.