Updated Trinity Study for 2021 -- thepoorswiss.com

[xbdb]

Very interesting data analysis with more recent data, longer horizons, and various withdrawal rates.

https://thepoorswiss.com/updated-trinity-study/

[vand]

It confirms what we already know but what might not be intuitively easy to understand:

- Stock-heavy portfolios will likely give you the best outcome
- Stock-heavy portfolios are the most likely to give you the worst outcome

These two statements not contradictory

[LD_TAndK]

Really interesting data and a well done blog, this conclusion seems like a stretch to me though

If you increase the simulation time to more than 30 years, a 4% withdrawal rate is no longer safe. With 50 years of retirement, you have a 90% chance of success with a 4% withdrawal rate at most. A withdrawal rate of around 3.5% would be safer for most people.

Over a 50 year timespan unforeseen life events will probably dominate the outcome of your finances, rather than consistent drawdown of capital. Focusing on the 10% chance of failure of an idealized drawdown seems to be missing the forest for the trees?

[dcheesi]

Really interesting data and a well done blog, this conclusion seems like a stretch to me though

If you increase the simulation time to more than 30 years, a 4% withdrawal rate is no longer safe. With 50 years of retirement, you have a 90% chance of success with a 4% withdrawal rate at most. A withdrawal rate of around 3.5% would be safer for most people.

Over a 50 year timespan unforeseen life events will probably dominate the outcome of your finances, rather than consistent drawdown of capital. Focusing on the 10% chance of failure of an idealized drawdown seems to be missing the forest for the trees?

True, OTOH, future upheavals that reduce the amount of reserve capital you need won't cause a problem if you have more, while those that increase the required capital will likely cause problems if you have less. So that seems like an additional argument in favor of a more conservative withdrawal strategy.

ETA: Of course you can't predict everything, and any additional reserve may still prove insufficient. But it makes sense to at least make sure you're fully covered for the scenarios you can plan for.

[TomTX]

It's a nice writeup. Hopefully the author will publish a followup with more modern high-success approaches (bond tent, variable withdrawal rate, etc)

Example: based on some online FIRE calculators, a VWR from 3.5% to 5% has a risk profile almost identical to a straight 3.5% withdrawal rate.

[Metalcat]

Really interesting data and a well done blog, this conclusion seems like a stretch to me though

If you increase the simulation time to more than 30 years, a 4% withdrawal rate is no longer safe. With 50 years of retirement, you have a 90% chance of success with a 4% withdrawal rate at most. A withdrawal rate of around 3.5% would be safer for most people.

Over a 50 year timespan unforeseen life events will probably dominate the outcome of your finances, rather than consistent drawdown of capital. Focusing on the 10% chance of failure of an idealized drawdown seems to be missing the forest for the trees?

True, OTOH, future upheavals that reduce the amount of reserve capital you need won't cause a problem if you have more, while those that increase the required capital will likely cause problems if you have less. So that seems like an additional argument in favor of a more conservative withdrawal strategy.

ETA: Of course you can't predict everything, and any additional reserve may still prove insufficient. But it makes sense to at least make sure you're fully covered for the scenarios you can plan for.

Yeah, but by that logic, no one should ever retire early.

[ender]

These types of "50 year retirements have 10% chance of running out of money" claims are always interesting.

In a 50 year timeline, more people will be dead than broke if only 10% of people have a retirement failure.

[maizefolk]

Yup. Assuming someone hits FIRE at 35, after 50 years more than half of women and more than two thirds of men will be dead.

To bring the risk of death over a 50 year time span UNDER 10% we have to assume FIRE at the age of 12 (for a woman) or 5 (for a man).

[norajean]

Unless you can guarantee you will be dead, isn’t it better to plan not to run out of money?

[Metalcat]

Unless you can guarantee you will be dead, isn’t it better to plan not to run out of money?

You can't actually plan that though.

If what you are saying is "isn't is better to save more rather than less?" Yeah, sure, but where do you draw the line?

I've said this many, many times, but you can't forget that the assumptions of FI simulators are pretty much nonsense. No one in the FI community is likely to spend precisely the same amount year over year plus inflation with total disregard for what the markets do. I just don't buy that.

Then there's the fact that the inputs into the calculator are completely made up. Is it a tight budget or a budget loaded with fat that the person is actually never likely to fully spend?
That will have far more impact than a .5% difference in calculated WR on very, very rough, made up numbers.

It's not that saving more isn't beneficial, obviously it is, but I'm personally not choosing to save more because of 90% vs 100% outputs from a simulator that don't even come close to comparing to the margins of error in the inputs in the first place.

[maizefolk]

Unless you can guarantee you will be dead, isn’t it better to plan not to run out of money?

The problem is that, by ignoring one risk, we end up overstating the other. If someone retiring at 35 has a 10% chance of running out of money by the time they reach 85 and a 2/3rd chance of dying before reaching 85, then they only have a roughly 3.3% chance* of living through the experience of running out of money which is what I, and I guess you as well, are actually worried about.

There is no 100% safe path. The earth could be hit by a giant asteroid tomorrow. Or the country in which you live could experience a communist and/or fascist revolution and all of your assets could be taken away. Or the country you live in could have a world war fought on its soil.** Or a garden variety economic collapse combined with lots of inflation like Argentina.

Since we can never rule out these bad things happening completely, the rough probability of a given bad thing happening matters a lot.

*That percentage is an over simplification since you could be dead by 85 but have still run out of money before you die. Fortunately calculators like @CCCA's actually can calculate the risk of running out of money before you die, rather than just assuming you'll live to a fixed age: https://engaging-data.com/will-money-last-retire-early/

**The main reason why FIRE success rates calculated using data from european countries which are not the UK or Switzerland tend to be lower than those calculated for the USA, Canada, UK, South Africa, Switzerland, New Zealand and Australia is the two world wars.

[ender]

It's all risk mitigation.

There's a massive thread here about "stop worrying about the 4% rule" talking about this.

[AO1FireTo]

These studies are good and it helps us plan.  I think the biggest risk running out of money, or death, it's the quality of your health.  If markets are bad, are you able to physically work, bring in some income to ride out any wild fluctuations in the market?  If you've optimized your spending to begin with, even part time work/business can significantly reduce any risk.  If you retire young enough, if markets are poor, instead of drawing down your investments, find a way to make an income for a year or two (or longer), to cover the bad years.  I think most people here, have a few backup plans should the market tank.  It's all about adjusting to what comes at you.

[shuffler]

Unless you can guarantee you will be dead, isn’t it better to plan not to run out of money?

Edna St. Vincent Millay - Midnight Oil

Cut if you will, with Sleep's Work's dull knife,
  Each day to half its length, my friend,–
The years that Time takes off my life
  He'll take from off the other end!

[maizefolk]

Unless you can guarantee you will be dead, isn’t it better to plan not to run out of money?

Edna St. Vincent Millay - Midnight Oil

Cut if you will, with Sleep's Work's dull knife,
  Each day to half its length, my friend,–
The years that Time takes off my life
  He'll take from off the other end!

Apologies in advance for the foam.

My first thought on reading this was "Wow! Midnight Oil had another song besides "Beds are Burning"?

[shuffler]

My first thought on reading this was "Wow! Midnight Oil had another song besides "Beds are Burning"?

Nope.  But they were big fans of Edna.

[Undecided]

Really interesting data and a well done blog, this conclusion seems like a stretch to me though

If you increase the simulation time to more than 30 years, a 4% withdrawal rate is no longer safe. With 50 years of retirement, you have a 90% chance of success with a 4% withdrawal rate at most. A withdrawal rate of around 3.5% would be safer for most people.

Over a 50 year timespan unforeseen life events will probably dominate the outcome of your finances, rather than consistent drawdown of capital. Focusing on the 10% chance of failure of an idealized drawdown seems to be missing the forest for the trees?

True, OTOH, future upheavals that reduce the amount of reserve capital you need won't cause a problem if you have more, while those that increase the required capital will likely cause problems if you have less. So that seems like an additional argument in favor of a more conservative withdrawal strategy.

ETA: Of course you can't predict everything, and any additional reserve may still prove insufficient. But it makes sense to at least make sure you're fully covered for the scenarios you can plan for.

Yeah, but by that logic, no one should ever retire early.

No one should ever retire *too* early. Finding that point is a challenge.

[mtnrider]

*That percentage is an over simplification since you could be dead by 85 but have still run out of money before you die. Fortunately calculators like @CCCA's actually can calculate the risk of running out of money before you die, rather than just assuming you'll live to a fixed age: https://engaging-data.com/will-money-last-retire-early/

This calculator should be at top of the forum!

I'd like to see it updated to include the chance that you'll be unable to do what you'd like.  For instance, one might make it to age 90, but odds are that for the last 15 years you'll only be able to reminisce about the things you were once able to do.

[norajean]

I'd like to see it updated to include the chance that you'll be unable to do what you'd like.  For instance, one might make it to age 90, but odds are that for the last 15 years you'll only be able to reminisce about the things you were once able to do.

Life ends at 75?  What?  I know scores of people who were vigorous well into their 90s.  Traveling, drinking, and taking on new adventures.  Unless you are hit with a major health issue, things just carry on.

[mtnrider]

I'd like to see it updated to include the chance that you'll be unable to do what you'd like.  For instance, one might make it to age 90, but odds are that for the last 15 years you'll only be able to reminisce about the things you were once able to do.

Life ends at 75?  What?  I know scores of people who were vigorous well into their 90s.  Traveling, drinking, and taking on new adventures.  Unless you are hit with a major health issue, things just carry on.

I don't know anyone in their 90s carrying on as they did at 40 (only 3 data points though).  After age 75 most people I know are moving quite a lot slower than they used to.  Parts are starting to wear out.  Age related disease gets you.  I think it's like 1:5 need some sort of mobility device by age 75 and it goes up steeply from there.  And IIRC, somewhere around 40% have Alzheimer's or dementia (from memory :)).

Everyone I know over age 85 are "contraindicated" for going on long flights due to various health problems (before covid!).

I watched my very athletic parents go from hiking, walking, and riding a bike about town to managing heart disease and joint replacements.  Now getting out of the house is a big adventure.

Heck, I'm only in my early 50s and I feel like I passed a bend point recently.  I need two or three pairs of glasses and my hearing in some frequencies is getting worse.  And I just can't push my heart rate up in time trial races like I did as a 20-something.

There are outliers, but it's a rare person hiking the PCT at age 75.

That's not to say life ends, or even that you will have less fun.  But you do need to redefine fun from what it was when you were younger.

I'm choosing to be realistic about my options, which feeds into why I want to get the bucket list items out of the way.  And thus FIRE.

[Metalcat]

I'd like to see it updated to include the chance that you'll be unable to do what you'd like.  For instance, one might make it to age 90, but odds are that for the last 15 years you'll only be able to reminisce about the things you were once able to do.

Life ends at 75?  What?  I know scores of people who were vigorous well into their 90s.  Traveling, drinking, and taking on new adventures.  Unless you are hit with a major health issue, things just carry on.

I don't know anyone in their 90s carrying on as they did at 40 (only 3 data points though).  After age 75 most people I know are moving quite a lot slower than they used to.  Parts are starting to wear out.  Age related disease gets you.  I think it's like 1:5 need some sort of mobility device by age 75 and it goes up steeply from there.  And IIRC, somewhere around 40% have Alzheimer's or dementia (from memory :)).

Everyone I know over age 85 are "contraindicated" for going on long flights due to various health problems (before covid!).

I watched my very athletic parents go from hiking, walking, and riding a bike about town to managing heart disease and joint replacements.  Now getting out of the house is a big adventure.

Heck, I'm only in my early 50s and I feel like I passed a bend point recently.  I need two or three pairs of glasses and my hearing in some frequencies is getting worse.  And I just can't push my heart rate up in time trial races like I did as a 20-something.

There are outliers, but it's a rare person hiking the PCT at age 75.

That's not to say life ends, or even that you will have less fun.  But you do need to redefine fun from what it was when you were younger.

I'm choosing to be realistic about my options, which feeds into why I want to get the bucket list items out of the way.  And thus FIRE.

As a health professional who has worked with a lot of seniors, it's really a bimodal distribution with the large peak being people who are quite infirm and the much smaller peak being very vital folks.

It's not common for 90+ people to be highly active and vital, but not as uncommon as most people assume.

The reason that most people think they never see vital 90+ folks is they don't appear to be 90+, and most people don't advertise their age.

You would be shocked at how many people you meet whom you assume are in their 70s who ate actually in their 90s and just never told you.

I know I wasn't very aware of this population until I started working a job where I had access to all of my patient's medical information and their age.

There really is a population that lives quite vital and healthy and die suddenly without much preceding degeneration.

[mtnrider]

As a health professional who has worked with a lot of seniors, it's really a bimodal distribution with the large peak being people who are quite infirm and the much smaller peak being very vital folks.

It's not common for 90+ people to be highly active and vital, but not as uncommon as most people assume.

The reason that most people think they never see vital 90+ folks is they don't appear to be 90+, and most people don't advertise their age.

You would be shocked at how many people you meet whom you assume are in their 70s who ate actually in their 90s and just never told you.

I know I wasn't very aware of this population until I started working a job where I had access to all of my patient's medical information and their age.

There really is a population that lives quite vital and healthy and die suddenly without much preceding degeneration.

Do you know if there have been studies on this?  I'd guess that it isn't betting odds - maybe 10-20% chance of remaining vital?

And it depends on one's definition of "vital".  For instance, I happen to know a 96 (!) year old woman who is quite spritely.  She's out walking around town and shopping, albeit with the help of a cane.  She does suffer from several minor illnesses (for instance, she needs a pacemaker), but you wouldn't know it from meeting her.  But she can't fly transcontinental due to those illnesses.

Looking at self-reported stats for hiking the Appalachian Trail and riding across the US, numbers fall off dramatically at age 75.

My original point was that I'd like to see the above Engaging Data tool updated with the chance of an illnesses that changes one's lifestyle.  I think that would be a useful tool for gauging when to FIRE, almost as useful as knowing the likelihood of death.  I don't think it's realistic for me to think that I'm as likely to be able to ride a TransAmerica route on my bike at 85 as I am at 55.  It could happen.  And I hope it does!  But I won't plan on it.

If I can get out and shop for groceries at 96.  I'll be happy! :)

[Metalcat]

As a health professional who has worked with a lot of seniors, it's really a bimodal distribution with the large peak being people who are quite infirm and the much smaller peak being very vital folks.

It's not common for 90+ people to be highly active and vital, but not as uncommon as most people assume.

The reason that most people think they never see vital 90+ folks is they don't appear to be 90+, and most people don't advertise their age.

You would be shocked at how many people you meet whom you assume are in their 70s who ate actually in their 90s and just never told you.

I know I wasn't very aware of this population until I started working a job where I had access to all of my patient's medical information and their age.

There really is a population that lives quite vital and healthy and die suddenly without much preceding degeneration.

Do you know if there have been studies on this?  I'd guess that it isn't betting odds - maybe 10-20% chance of remaining vital?

And it depends on one's definition of "vital".  For instance, I happen to know a 96 (!) year old woman who is quite spritely.  She's out walking around town and shopping, albeit with the help of a cane.  She does suffer from several minor illnesses (for instance, she needs a pacemaker), but you wouldn't know it from meeting her.  But she can't fly transcontinental due to those illnesses.

Looking at self-reported stats for hiking the Appalachian Trail and riding across the US, numbers fall off dramatically at age 75.

My original point was that I'd like to see the above Engaging Data tool updated with the chance of an illnesses that changes one's lifestyle.  I think that would be a useful tool for gauging when to FIRE, almost as useful as knowing the likelihood of death.  I don't think it's realistic for me to think that I'm as likely to be able to ride a TransAmerica route on my bike at 85 as I am at 55.  It could happen.  And I hope it does!  But I won't plan on it.

If I can get out and shop for groceries at 96.  I'll be happy! :)

Well, that's what I mean by bimodal distribution with the small peak being really vital 90+ year olds. There are few of them, but they do exist. And I'm not talking "doing well for 90+", I mean that they're vital for being seniors, period. No medications, no chronic illness other than maybe a bit of arthritis.

I was on the phone this morning with a friend who lives with his 95 year old mother and they were rebuilding furniture together. She's literally in better shape and more active than half of the middle aged people I know.

Yes that's an anecdote but my entire point is anecdotal. My clinical experience is that there is a small population of these nearly 100 year old people who are in great shape.

The majority are in terrible shape and very unhealthy and there aren't a lot of people in between.

[slappy]

As a health professional who has worked with a lot of seniors, it's really a bimodal distribution with the large peak being people who are quite infirm and the much smaller peak being very vital folks.

It's not common for 90+ people to be highly active and vital, but not as uncommon as most people assume.

The reason that most people think they never see vital 90+ folks is they don't appear to be 90+, and most people don't advertise their age.

You would be shocked at how many people you meet whom you assume are in their 70s who ate actually in their 90s and just never told you.

I know I wasn't very aware of this population until I started working a job where I had access to all of my patient's medical information and their age.

There really is a population that lives quite vital and healthy and die suddenly without much preceding degeneration.

Do you know if there have been studies on this?  I'd guess that it isn't betting odds - maybe 10-20% chance of remaining vital?

And it depends on one's definition of "vital".  For instance, I happen to know a 96 (!) year old woman who is quite spritely.  She's out walking around town and shopping, albeit with the help of a cane.  She does suffer from several minor illnesses (for instance, she needs a pacemaker), but you wouldn't know it from meeting her.  But she can't fly transcontinental due to those illnesses.

Looking at self-reported stats for hiking the Appalachian Trail and riding across the US, numbers fall off dramatically at age 75.

My original point was that I'd like to see the above Engaging Data tool updated with the chance of an illnesses that changes one's lifestyle.  I think that would be a useful tool for gauging when to FIRE, almost as useful as knowing the likelihood of death.  I don't think it's realistic for me to think that I'm as likely to be able to ride a TransAmerica route on my bike at 85 as I am at 55.  It could happen.  And I hope it does!  But I won't plan on it.

If I can get out and shop for groceries at 96.  I'll be happy! :)

Out of curiosity, why is hiking the appalachian trail the benchmark you are using for vitality?

[TomTX]

Yes that's an anecdote but my entire point is anecdotal. My clinical experience is that there is a small population of these nearly 100 year old people who are in great shape.

The majority are in terrible shape and very unhealthy and there aren't a lot of people in between.

Not to mention the rather large peak of "already dead" - a portion which rarely show up to interact/socialize and thus make mindshare.

[mtnrider]

Out of curiosity, why is hiking the appalachian trail the benchmark you are using for vitality?

Because it's all about ME, and MY bucket list!  :)

Originally that included

• hang gliding along the US/Canada line
• canoeing the Moisie River
• kayaking the Alaskan or Canadian coast
• hiking one or more of the long trails
• bicycling across the US
• circumnavigating the globe on a smallish sailboat

I happen to know a handful of people who have done some of those big adventures and are now 75+.  They all tell me to take adventures while I'm still relatively young, that at their age they can only relate their fun adventures, not participate.

As a younger man, if I had a graph similar to the Rich, Broke, or Dead app, but one that also included how likely I'd be to able to actually do an adventure, I might have been able to arrange my priorities differently.  Naive, I know, but at age 25 I figured I could work until 67 and then start on the bucket list.

[Metalcat]

Yes that's an anecdote but my entire point is anecdotal. My clinical experience is that there is a small population of these nearly 100 year old people who are in great shape.

The majority are in terrible shape and very unhealthy and there aren't a lot of people in between.

Not to mention the rather large peak of "already dead" - a portion which rarely show up to interact/socialize and thus make mindshare.

I didn't feel the need to mention them since they had already been addressed in the thread.

[maizefolk]

As a younger man, if I had a graph similar to the Rich, Broke, or Dead app, but one that also included how likely I'd be to able to actually do an adventure, I might have been able to arrange my priorities differently.  Naive, I know, but at age 25 I figured I could work until 67 and then start on the bucket list.

If someone can find data on the proportion of people who are in good health at each age exists out there, I bet it could be incorporated into the calculator.

I think the big challenges are:

1) Defining what constitutes "good health." For one person it means the ability to hike for a week and sleep on the ground every night, for another it might simply mean being able to walk unassisted, for a third it might boil down to not having dementia and worrying much less about physical capabilities.

2) Getting data at high enough resolution (would need it on a year by year basis, a lot of the dataset's I've found aggregate people into cohorts of a decade or more).

3) The "but I'm extra healthy and eat lots of salads" argument which already comes up with the death rate. Although I see the calculator has added life expectancy data for three levels of health now. Constantly getting more features.

[NorthernMonkey]

Studies like this do not take account that someone who is FIRE is unlikely to be totally oblivious to market forces.

Just like a huge crash has implications for those working, as they dust of their resume and tighten their belts, it should also have some, but lesser implications for those FIRE. Im sure everyone who has FIREd has a few months of expenses as cash, or even access to credit cards.

If there has been a huge 40% dip in the markets, then knowing this is the time to spend cash, or on CCs, instead of selling stock at its lowest should be something most FIREees are more than aware and capable of calculating.

4%, but with some minor controls seems sensible. To get from 4% to 3.5% means saving for an extra 12%. All that time slaving away that could be avoided.

[Zinsch]

To me, the biggest question is whether past returns are sufficiently predictive of future ones. We are in a low interest rate environment right now, which is why stock valuations are high. Expected future equity returns are probably lower than the historical average. This would also impact the safe withdrawal rate.

[wageslave23]

To me, the biggest question is whether past returns are sufficiently predictive of future ones. We are in a low interest rate environment right now, which is why stock valuations are high. Expected future equity returns are probably lower than the historical average. This would also impact the safe withdrawal rate.

Exactly.  I reject the premise that the last 3 30 yr time periods have much predictive value for the next 30 years.  Let alone the next 2 30 yr time periods.  I am going to loosely base my FIRE on historical returns because thats all we have to go on.  But to get too detailed in analysis of historical returns and predicting health and death rates down to .5 percentage points definitely is missing the forest for the trees.  Build up enough nest egg that historical would be about enough, have some backup plans and then go for it.  The key is that if you are happy living on almost nothing then there are a lot of safety nets to fall back on.  The old people, who's lives i would like to emulate, spend very little money at all.

[ender]

To me, the biggest question is whether past returns are sufficiently predictive of future ones. We are in a low interest rate environment right now, which is why stock valuations are high. Expected future equity returns are probably lower than the historical average. This would also impact the safe withdrawal rate.

I personally think that it's actually the opposite impact - I think it's likely that future returns, while perhaps lower, will be more stable.

The government has shown an increasing interest in trying to stabilize the economy in the last few decades as compared with a huge percentage of the time that FIRE calcs incorporate.

If I could FIRE and guarantee 2-3% after inflation returns, every year, with no variation, that's almost perfect for FIRE on 4%. You eventually draw it down but it'd take a long time.

Rarely does overall return % cause FIRE scenarios to fail. Sequence of return risk (along with inflation) are the big problems when backtesting.