MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by skierincolorado »

HomerJ wrote: Tue Aug 03, 2021 11:54 pm
skierincolorado wrote: Tue Aug 03, 2021 11:14 pmSurely the odds of such scenarios are heightened when the mean expectation is 3-4% instead of 7%.
That's just it. We don't know. It's not an independent variable. It may not be a normal distribution that just evenly moves all possible returns to the left along with the mean.

7% mean may indicate 1%-13% (the range is so large the prediction is really almost meaningless, but let's ignore that).

4% mean doesn't necessarily automatically translate to -2% to 10%.

We don't know enough. There are too many variables. Not enough data points. Human emotions are involved, and governments can and do step in.
Exactly. We don't know. All we know is that the mean expectation is lower. In order to not have lower retirement expectations, you would have to ASSUME that the standard deviation has been cut in half. And that's something we don't know.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by skierincolorado »

YRT70 wrote: Wed Aug 04, 2021 12:58 am
skierincolorado wrote: Tue Aug 03, 2021 3:46 pm
HomerJ wrote: Tue Aug 03, 2021 2:52 pm
nigel_ht wrote: Tue Aug 03, 2021 2:06 pm
young-ish wrote: Tue Aug 03, 2021 2:02 pm

How did you arrive at 3.25%? Is that a constant WD % or a starting with inflation adjusted increases? What asset allocation supports this WDR?
Image

https://earlyretirementnow.com/2016/12/ ... t-1-intro/
So that chart shows 4% as fine for 30 year retirements. I'm not sure where someone gets the idea that it is pretty well established that 4% doesn't work for 30 years.
The person who made that chart (ERN) would tell you in his numerous other blog posts that 4% is not sufficient for over 30 years due to current low interest rates and correspondingly high valuations.
Got a link to where he says that?
https://earlyretirementnow.com/2016/12/ ... valuation/
YRT70
Posts: 1289
Joined: Sat Apr 27, 2019 8:51 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by YRT70 »

skierincolorado wrote: Wed Aug 04, 2021 9:15 am
YRT70 wrote: Wed Aug 04, 2021 12:58 am
skierincolorado wrote: Tue Aug 03, 2021 3:46 pm
HomerJ wrote: Tue Aug 03, 2021 2:52 pm
So that chart shows 4% as fine for 30 year retirements. I'm not sure where someone gets the idea that it is pretty well established that 4% doesn't work for 30 years.
The person who made that chart (ERN) would tell you in his numerous other blog posts that 4% is not sufficient for over 30 years due to current low interest rates and correspondingly high valuations.
Got a link to where he says that?
https://earlyretirementnow.com/2016/12/ ... valuation/
I also missed the word "over". I was aware that 4% is not a good idea for long retirements.
User avatar
Ice-9
Posts: 1579
Joined: Wed Oct 15, 2008 12:40 pm
Location: MD

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by Ice-9 »

Grifin wrote: Tue Aug 03, 2021 12:43 pm These figures differ greatly from many popular retirement modelling tools, ie. FireCalc. In true MW fashion, there is no real detail offered to understand the methods, data and assumptions for the analysis.
I just wanted to note that FireCalc uses actual historical returns, and this study apparently assumed worse than historical.

In FireCalc, you can account for worse than historical returns by adding to the expense ratio on the Your Portfolio tab. For example, if your actual weighted average expense ratio for your portfolio is 0.08% and you expect returns 2% worse than historical going forward, you can enter your expense ratio as 2.08% to simulate that.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by HomerJ »

skierincolorado wrote: Wed Aug 04, 2021 9:11 am
HomerJ wrote: Tue Aug 03, 2021 11:54 pm
skierincolorado wrote: Tue Aug 03, 2021 11:14 pmSurely the odds of such scenarios are heightened when the mean expectation is 3-4% instead of 7%.
That's just it. We don't know. It's not an independent variable. It may not be a normal distribution that just evenly moves all possible returns to the left along with the mean.

7% mean may indicate 1%-13% (the range is so large the prediction is really almost meaningless, but let's ignore that).

4% mean doesn't necessarily automatically translate to -2% to 10%.

We don't know enough. There are too many variables. Not enough data points. Human emotions are involved, and governments can and do step in.
Exactly. We don't know. All we know is that the mean expectation is lower. In order to not have lower retirement expectations, you would have to ASSUME that the standard deviation has been cut in half. And that's something we don't know.
No, because the original retirement expectations were not based on the mean expectation of the past.

4% is based on the worst cases of the past, not the average years of the past.

4% is already based on low returns. So one stating "Hey, it looks like like we're likely to get low returns for the next 10 years" just means we can say "Whew, good thing we were already using a plan that works with low returns".

The one thing we DO know is that 4% hasn't failed yet (okay 3.8% in 1966)... And during some pretty tough financial times. The next 30 years would have be WORSE than the Great Depression with 25% unemployment or the 1966-1982 period with 16 years of 0% real stock growth and negative real bond returns.

Those were pretty bad times, and 4% worked. How do you calculate the odds that the next 30 years will be worse than the Great Depression?

BUT... we are talking about 30+ year retirements and FIRE, and yes 4% was not good enough in the past for some periods of 30 PLUS years.

So, again, my apologies for arguing so strongly when we weren't talking about 30 year retirements, but instead talking about 40, 50 year retirements. I agree that a perpetual rate of 3% would be more safe for such long retirements.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
User avatar
tennisplyr
Posts: 3703
Joined: Tue Jan 28, 2014 12:53 pm
Location: Sarasota, FL

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by tennisplyr »

What about the folks who didn’t make it to 50 years…were they married, did they work in retirement, what were their expenses, where did they live, did they own their own home, did they have children, etc, etc. I have control of my financial destiny and I will make it.
“Those who move forward with a happy spirit will find that things always work out.” -Retired 13 years 😀
Angst
Posts: 2968
Joined: Sat Jun 09, 2007 11:31 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by Angst »

livesoft wrote: Tue Aug 03, 2021 12:49 pm Early Retirement Now has explored all this in great detail, so I think it is not a bad thing that others are finding not dissimilar results.
https://earlyretirementnow.com/safe-wit ... te-series/
I too think it's a good thing that others are finding similar results!
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by skierincolorado »

HomerJ wrote: Wed Aug 04, 2021 11:09 am
skierincolorado wrote: Wed Aug 04, 2021 9:11 am
HomerJ wrote: Tue Aug 03, 2021 11:54 pm
skierincolorado wrote: Tue Aug 03, 2021 11:14 pmSurely the odds of such scenarios are heightened when the mean expectation is 3-4% instead of 7%.
That's just it. We don't know. It's not an independent variable. It may not be a normal distribution that just evenly moves all possible returns to the left along with the mean.

7% mean may indicate 1%-13% (the range is so large the prediction is really almost meaningless, but let's ignore that).

4% mean doesn't necessarily automatically translate to -2% to 10%.

We don't know enough. There are too many variables. Not enough data points. Human emotions are involved, and governments can and do step in.
Exactly. We don't know. All we know is that the mean expectation is lower. In order to not have lower retirement expectations, you would have to ASSUME that the standard deviation has been cut in half. And that's something we don't know.
No, because the original retirement expectations were not based on the mean expectation of the past.

4% is based on the worst cases of the past, not the average years of the past.

4% is already based on low returns. So one stating "Hey, it looks like like we're likely to get low returns for the next 10 years" just means we can say "Whew, good thing we were already using a plan that works with low returns".

The one thing we DO know is that 4% hasn't failed yet (okay 3.8% in 1966)... And during some pretty tough financial times. The next 30 years would have be WORSE than the Great Depression with 25% unemployment or the 1966-1982 period with 16 years of 0% real stock growth and negative real bond returns.

Those were pretty bad times, and 4% worked. How do you calculate the odds that the next 30 years will be worse than the Great Depression?

BUT... we are talking about 30+ year retirements and FIRE, and yes 4% was not good enough in the past for some periods of 30 PLUS years.

So, again, my apologies for arguing so strongly when we weren't talking about 30 year retirements, but instead talking about 40, 50 year retirements. I agree that a perpetual rate of 3% would be more safe for such long retirements.
And the Vanguard, nor the ERN series, are not based on the mean expectation being 1929 either. The mean expectation is only somewhat lower than in the past - and with very strong evidence that expectations should be reduced. If the mean is reduced it is just basic common sense that the worst cases and the entire distribution of outcomes is reduced as well. Unless you have a strong argument that the variance is reduced as well. But I've heard no such argument other than a guess that the U.S. govt might do something. But the U.S. govt couldn't/didn't stop 1929 or 1966 did they?

FWIW the ERN spreadsheet spits out more like 3.7-3.8% than the 3.25% I suggested. We're not arguing over a lot here. There are others here that will tell you that it needs to be 3% flat or less which doesn't make sense because 3% would last 30 years if I could just maintain the real value of my money - in TIPS for example. And surely diversifying with a little domestic and intl equities could only help over 30 years.

It doesn't make sense to assume that the odds of 1966 are the same in the future as they were historically when we know the future doesn't look as bright. Just generally speaking this requires at least a shred more caution but of course that can be hashed out mathematically as well.

Personally, I'd probably still be fine going for 4% if I knew I had the ability to reduce spending if necessary. But I'd be very sanguine about that possibility being higher than it has been historically. Somebody with 25x expenses isn't going to starve if they have some common sense. The plan needs to be reevaluated each year. Of course if this was 1985 I'd probably have been fine going with 4.5% if I had flexibility in expenses.
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by alex_686 »

LookingForward wrote: Tue Aug 03, 2021 8:42 pm
alex_686 wrote: Tue Aug 03, 2021 12:57 pm One can make relatively decent 10 year forecasts.
Is this true?

Can one predict the most recent 10 years from the previous data? And so on?

I’ve been seeing dire predictions for the market and the economy, supposedly based on hard data, since I started paying attention, but that was only a dozen years ago.
Yes. But not on the past 10 year's worth of data. More on that latter.

While it is hard to see what is going to happen the sort run, it is easier to see the big picture.

I don't want to oversell this or undersell this. You can look at the grand sweep of things. You can look at demographics and get a idea of savings rates and investment demands. GNP is a strong indirect driver of stock returns and you can see the grand sweeps before they are coming. It takes time for technology and productivity gains to ramp up. People need to be trained, factories built. etc. Or you can look at long term earnings growth. etc. There are a dozen different ways to come up with estimates. CAPE 10 is one. See below for more.

Are there wide error bars? Sure. Does "Nobody Know Nothing", No. It is a coinvent dodge for thinking about large difficult questions with a high degree of uncertainty. However you can't really put together a ISP or AA unless you have some market expectations. You are going to get what the market gives. But there is a trade off between current consumption and saving for future consumption. You have to have some framework to figure out how to balance these things out.

So, 2 different scenarios.

25 years ago, expected inflation was about 3%, 10 Year Treasuries were 5%, so a 2% expected real return on Treasuries. The expected long term return on stocks was around 10%, so there was a 5% equity risk premium.

Today expected inflation is around 2%, 10 year Treasuries around 2%, so 0% expected real retunes. Long term return on stocks is around 7% to 10% depending on how you read the numbers. So that is a 5% to 8% equity risk premium. So, same equity return with more risk.

So, let us talk about risk and prior dismal returns. I have been one of those nay-sayers.

The bond side is easier. Let us set up the context. For the past 10 years we have been in a great bond rally. I though 10 year Treasuries were about as low as they could go at a mid 3% yield. I was wrong. They have fallen to 2%. Can they fall any lower? Probably not. This limits any gains there. Can they go up? Yes. So the prices moves are not symmetrical this has a impact on the expected risk and return profile. The future prospects are bleak. Low expected returns, high risk.

Now, is this market timing? No. Market timing is reducing or shorting Treasuries because you think they will go down shortly. I don't know. While they really can't go up they could go sideways for years.

So, going back to predicting dismal returns.

Returns for the past 10 years have not been driven by higher returns on assets. i.e., not by high bond yields or high E/P ratios (the inverse of the P/E ratio) or high dividend yields. Rather, assets have had a one time boost as these ratios have fallen. These ratios can't continue to fall forever. We can debate where the limit is, but there is a limit and we are getting close to what that limit is.

So, given the current ratio levels, my market expectations are low.

Now, can we predict the future 10 years from the prior 10 years? Yes and No.

I like the CAPE 10 model, and promote it on Bogleheads. Yes, it has huge error bands. But we are talking about a very simple model. The theory is decent. It makes reasonable decent predictions that are actionable. You can squeeze the formula and its underlying assumptions on a 3x5 card. It is objective and mechanical. All of the inputs are publicly available.

Are there issues with the model? Yeah - mainly because it is so simple. For example, you can drastically improve its predictive power if you adjust the earnings to reflect the accounting changes on how earnings are calculated and reported during that 10 year period. However, this is complex, nuanced, requires some subjective decisions, and requires proprietary data. So, this is a minus for a public forum.

And on CAPE 10 specifically, the COVID pandemic has broken some of its assumptions. It assumes that the next 10 years will be like the last 10 years. CAPE 10 is basically a noise reduction algorithm so you can see the long term trends in earnings. In the short run earnings can be manipulated with one-off charges. COVID is not a one-off charge to the earnings. Rather it is a break in the continuous data. You are going to have a issue with any model if you try to span a period with this data. The next 10 years won't be like the past pandemic period. Adjustments will need to be made.

But there are multiple models out there. You can use intelligence, logic, and wisdom to come up with a answer that works for you.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
marcopolo
Posts: 8445
Joined: Sat Dec 03, 2016 9:22 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by marcopolo »

alex_686 wrote: Wed Aug 04, 2021 1:57 pm
LookingForward wrote: Tue Aug 03, 2021 8:42 pm
alex_686 wrote: Tue Aug 03, 2021 12:57 pm One can make relatively decent 10 year forecasts.
Is this true?

Can one predict the most recent 10 years from the previous data? And so on?

I’ve been seeing dire predictions for the market and the economy, supposedly based on hard data, since I started paying attention, but that was only a dozen years ago.
Yes. But not on the past 10 year's worth of data. More on that latter.

While it is hard to see what is going to happen the sort run, it is easier to see the big picture.

I don't want to oversell this or undersell this. You can look at the grand sweep of things. You can look at demographics and get a idea of savings rates and investment demands. GNP is a strong indirect driver of stock returns and you can see the grand sweeps before they are coming. It takes time for technology and productivity gains to ramp up. People need to be trained, factories built. etc. Or you can look at long term earnings growth. etc. There are a dozen different ways to come up with estimates. CAPE 10 is one. See below for more.

Are there wide error bars? Sure. Does "Nobody Know Nothing", No. It is a coinvent dodge for thinking about large difficult questions with a high degree of uncertainty. However you can't really put together a ISP or AA unless you have some market expectations. You are going to get what the market gives. But there is a trade off between current consumption and saving for future consumption. You have to have some framework to figure out how to balance these things out.

So, 2 different scenarios.

25 years ago, expected inflation was about 3%, 10 Year Treasuries were 5%, so a 2% expected real return on Treasuries. The expected long term return on stocks was around 10%, so there was a 5% equity risk premium.

Today expected inflation is around 2%, 10 year Treasuries around 2%, so 0% expected real retunes. Long term return on stocks is around 7% to 10% depending on how you read the numbers. So that is a 5% to 8% equity risk premium. So, same equity return with more risk.

So, let us talk about risk and prior dismal returns. I have been one of those nay-sayers.

The bond side is easier. Let us set up the context. For the past 10 years we have been in a great bond rally. I though 10 year Treasuries were about as low as they could go at a mid 3% yield. I was wrong. They have fallen to 2%. Can they fall any lower? Probably not. This limits any gains there. Can they go up? Yes. So the prices moves are not symmetrical this has a impact on the expected risk and return profile. The future prospects are bleak. Low expected returns, high risk.

Now, is this market timing? No. Market timing is reducing or shorting Treasuries because you think they will go down shortly. I don't know. While they really can't go up they could go sideways for years.

So, going back to predicting dismal returns.

Returns for the past 10 years have not been driven by higher returns on assets. i.e., not by high bond yields or high E/P ratios (the inverse of the P/E ratio) or high dividend yields. Rather, assets have had a one time boost as these ratios have fallen. These ratios can't continue to fall forever. We can debate where the limit is, but there is a limit and we are getting close to what that limit is.

So, given the current ratio levels, my market expectations are low.

Now, can we predict the future 10 years from the prior 10 years? Yes and No.

I like the CAPE 10 model, and promote it on Bogleheads. Yes, it has huge error bands. But we are talking about a very simple model. The theory is decent. It makes reasonable decent predictions that are actionable. You can squeeze the formula and its underlying assumptions on a 3x5 card. It is objective and mechanical. All of the inputs are publicly available.

Are there issues with the model? Yeah - mainly because it is so simple. For example, you can drastically improve its predictive power if you adjust the earnings to reflect the accounting changes on how earnings are calculated and reported during that 10 year period. However, this is complex, nuanced, requires some subjective decisions, and requires proprietary data. So, this is a minus for a public forum.

And on CAPE 10 specifically, the COVID pandemic has broken some of its assumptions. It assumes that the next 10 years will be like the last 10 years. CAPE 10 is basically a noise reduction algorithm so you can see the long term trends in earnings. In the short run earnings can be manipulated with one-off charges. COVID is not a one-off charge to the earnings. Rather it is a break in the continuous data. You are going to have a issue with any model if you try to span a period with this data. The next 10 years won't be like the past pandemic period. Adjustments will need to be made.

But there are multiple models out there. You can use intelligence, logic, and wisdom to come up with a answer that works for you.
I don't have nearly the faith you have in these models. I think past history (how poorly most models have predicted market behavior) supports my skepticism.

I do agree with your last paragraph, there are so many different models out there, one can construct whatever answer they want, which I suspect is what most people do to support there own preconceived biases.
Once in a while you get shown the light, in the strangest of places if you look at it right.
flyingaway
Posts: 3908
Joined: Fri Jan 17, 2014 9:19 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by flyingaway »

Once in a short while, there will be a repeated discussion about the 4% rule.

Maybe in the Miriam2's annual Roll Call for the Retirement Class thread, we should ask potential retirees to list their planned withdrawal rates.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by HomerJ »

skierincolorado wrote: Wed Aug 04, 2021 1:36 pmIf the mean is reduced it is just basic common sense that the worst cases and the entire distribution of outcomes is reduced as well.
That does seem like common sense, doesn't it?

For an independent variable with a purely random normal distribution, sure...

But there are a ton of other variables. Do you really think annual returns are purely random events? Do last year returns have an effect on this year's returns? Human emotions at play here, Fed decisions, media, government actions affecting the economy.

Low returns aren't enough to break 4%. High inflation is required. Is high inflation a purely random event? Is there an government agency that monitors inflation and has tools to combat it?
Unless you have a strong argument that the variance is reduced as well. But I've heard no such argument other than a guess that the U.S. govt might do something. But the U.S. govt couldn't/didn't stop 1929 or 1966 did they?
Of course I have a strong argument. The government did everything wrong in 1929. Ben Bernanke, the Fed Chairman in 2008, had studied the Great Depression in depth, and did everything he could to stop a new global depression.

You mention 1966... Inflation was stopped in 1982, and 4% survived because of government action by raising interest rates.

Am I stating that we will never see financial difficulties again because the government will always protect us... No way!

But I don't see returns as a purely random normal distribution. If the tail end edges down into negative territory because the mean shifts left, there will be plenty of resistance to keep actual returns positive. "It's the economy, stupid" is a lesson every politician has learned.
FWIW the ERN spreadsheet spits out more like 3.7-3.8% than the 3.25% I suggested. We're not arguing over a lot here. There are others here that will tell you that it needs to be 3% flat or less which doesn't make sense because 3% would last 30 years if I could just maintain the real value of my money - in TIPS for example. And surely diversifying with a little domestic and intl equities could only help over 30 years.

It doesn't make sense to assume that the odds of 1966 are the same in the future as they were historically when we know the future doesn't look as bright. Just generally speaking this requires at least a shred more caution but of course that can be hashed out mathematically as well.

Personally, I'd probably still be fine going for 4% if I knew I had the ability to reduce spending if necessary. But I'd be very sanguine about that possibility being higher than it has been historically. Somebody with 25x expenses isn't going to starve if they have some common sense. The plan needs to be reevaluated each year. Of course if this was 1985 I'd probably have been fine going with 4.5% if I had flexibility in expenses.
I agree that we're really not arguing over much. In practical terms, 3.7% and 4% (with discretionary expenses) aren't much different.
Last edited by HomerJ on Wed Aug 04, 2021 2:55 pm, edited 2 times in total.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
FactualFran
Posts: 2776
Joined: Sat Feb 21, 2015 1:29 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by FactualFran »

YRT70 wrote: Wed Aug 04, 2021 10:42 am I also missed the word "over". I was aware that 4% is not a good idea for long retirements.
The word "over" in the phrase "sufficient for over 30 years" has multiple possible interpretations. In one sense, it means sufficient for periods of 30 years. In another sense, it means was sufficient for periods of more than 30 years.

Here the context is withdrawals rates. A condition commonly used by withdrawal rate studies is an approach working for periods of at least 30 years. The first interpretation of "over" seems more natural in that context.
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by alex_686 »

marcopolo wrote: Wed Aug 04, 2021 2:13 pm I don't have nearly the faith you have in these models. I think past history (how poorly most models have predicted market behavior) supports my skepticism.

I do agree with your last paragraph, there are so many different models out there, one can construct whatever answer they want, which I suspect is what most people do to support there own preconceived biases.
Can you explain why you prefer past history? So, my position: Models bad, history worse.

Historical models offer very little information. Worse, most people don't have a firm grasp on statistics or the underlying data. They come away worse from the experience.

There are "secular periods" (the technical economic term) in economics and investing. 5 to 15 years are common but there are examples far outside these boundaries. This period will have a mean reverting characteristic. Stock and bond returns, volatility, and correlations. Then at some random interval the economic and investing system will shift to a new state with new characteristics.

You have to keep your analysis to within that period. If you cross periods you are probably breaking your statistical assumption. You need to reach for more complex methods like a "Jump Diffusion Model".

You can't use analysis from a prior secular period for the current period. I mean, we are talking about different system. They are going to act differently.

However, you can validate a model using data from prior periods. Then use current data to run your model.

So, dig through the academic data and you will see that the models have preformed far better than historical analysis for long term planning.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by HomerJ »

alex_686 wrote: Wed Aug 04, 2021 2:55 pm
marcopolo wrote: Wed Aug 04, 2021 2:13 pm I don't have nearly the faith you have in these models. I think past history (how poorly most models have predicted market behavior) supports my skepticism.

I do agree with your last paragraph, there are so many different models out there, one can construct whatever answer they want, which I suspect is what most people do to support there own preconceived biases.
Can you explain why you prefer past history? So, my position: Models bad, history worse.

Historical models offer very little information. Worse, most people don't have a firm grasp on statistics or the underlying data. They come away worse from the experience.

There are "secular periods" (the technical economic term) in economics and investing. 5 to 15 years are common but there are examples far outside these boundaries. This period will have a mean reverting characteristic. Stock and bond returns, volatility, and correlations. Then at some random interval the economic and investing system will shift to a new state with new characteristics.

You have to keep your analysis to within that period. If you cross periods you are probably breaking your statistical assumption. You need to reach for more complex methods like a "Jump Diffusion Model".

You can't use analysis from a prior secular period for the current period. I mean, we are talking about different system. They are going to act differently.

However, you can validate a model using data from prior periods. Then use current data to run your model.

So, dig through the academic data and you will see that the models have preformed far better than historical analysis for long term planning.
Historical models when you are just looking at worst case are extremely useful. You are technically making a prediction, but a super conservative very broad one.

Imagine building a house near a river. You check the historical record of floods in the past 100 years.

Say the worst case flood ever (so far!) was 20 feet high, and that only happened once.

Building your house 22 feet above the river is a pretty conservative choice. The historical data is indeed useful and it doesn't require a model that accurately predicts how often floods will hit 8 feet or 12 feet or 15 feet. I don't care.

Building your house 17 feet above the river is a lot more risky, and you'll need a better model of what conditions caused 17-20 foot floods in the past, and how likely it is to happen in the next 10-20 years.

Sure, some variables may change, and maybe there will be a 23-foot flood in the future. That's always possible.

But if you come along with a model that says "hey, it looks like we're due for some higher than average floods going forward", I would say "Well, good thing I picked a spot higher than the worst flood so far".
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
marcopolo
Posts: 8445
Joined: Sat Dec 03, 2016 9:22 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by marcopolo »

alex_686 wrote: Wed Aug 04, 2021 2:55 pm
marcopolo wrote: Wed Aug 04, 2021 2:13 pm I don't have nearly the faith you have in these models. I think past history (how poorly most models have predicted market behavior) supports my skepticism.

I do agree with your last paragraph, there are so many different models out there, one can construct whatever answer they want, which I suspect is what most people do to support there own preconceived biases.
Can you explain why you prefer past history? So, my position: Models bad, history worse.

Historical models offer very little information. Worse, most people don't have a firm grasp on statistics or the underlying data. They come away worse from the experience.

There are "secular periods" (the technical economic term) in economics and investing. 5 to 15 years are common but there are examples far outside these boundaries. This period will have a mean reverting characteristic. Stock and bond returns, volatility, and correlations. Then at some random interval the economic and investing system will shift to a new state with new characteristics.

You have to keep your analysis to within that period. If you cross periods you are probably breaking your statistical assumption. You need to reach for more complex methods like a "Jump Diffusion Model".

You can't use analysis from a prior secular period for the current period. I mean, we are talking about different system. They are going to act differently.

However, you can validate a model using data from prior periods. Then use current data to run your model.

So, dig through the academic data and you will see that the models have preformed far better than historical analysis for long term planning.
What you are describing can be reasonably summed up as non-stationarity. The problem is most models assume some level of stationarity. CAPE10 is a classic example. It was mined from data prior to 1994. It has been a pretty poor predictor of markets ever since. I have shown in another thread at least one seemingly abrupt change in the relationship between CAPE10 ans subsequent 10 yr performance.

My comment about history supporting my skepticism of the models was not meant to convey that I "trust" history to repeat it self, but rather to point at the weakness of the models.
Once in a while you get shown the light, in the strangest of places if you look at it right.
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by nigel_ht »

skierincolorado wrote: Wed Aug 04, 2021 1:36 pm But I've heard no such argument other than a guess that the U.S. govt might do something. But the U.S. govt couldn't/didn't stop 1929 or 1966 did they?
The fed learned from 1929 given it made the wrong move. Whether unfettered QE is the right move is somewhat TBD but worked in the GFC.

1970s taught “inflation BAAAAD”. Whether the fed can really stop inflation is also TBD…but it was partially caused by the govt trying to drop unemployment by providing easy money. So far the jury is out whether the current inflationary pressures are temporary or not.

So we’re not likely repeat those particular mistakes but invent some new epic mistake. Predicting that will be challenging.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by skierincolorado »

HomerJ wrote: Wed Aug 04, 2021 2:42 pm
skierincolorado wrote: Wed Aug 04, 2021 1:36 pmIf the mean is reduced it is just basic common sense that the worst cases and the entire distribution of outcomes is reduced as well.
That does seem like common sense, doesn't it?

For an independent variable with a purely random normal distribution, sure...

But there are a ton of other variables. Do you really think annual returns are purely random events? Do last year returns have an effect on this year's returns? Human emotions at play here, Fed decisions, media, government actions affecting the economy.

Low returns aren't enough to break 4%. High inflation is required. Is high inflation a purely random event? Is there an government agency that monitors inflation and has tools to combat it?
Unless you have a strong argument that the variance is reduced as well. But I've heard no such argument other than a guess that the U.S. govt might do something. But the U.S. govt couldn't/didn't stop 1929 or 1966 did they?
Of course I have a strong argument. The government did everything wrong in 1929. Ben Bernanke, the Fed Chairman in 2008, had studied the Great Depression in depth, and did everything he could to stop a new global depression.

You mention 1966... Inflation was stopped in 1982, and 4% survived because of government action by raising interest rates.

Am I stating that we will never see financial difficulties again because the government will always protect us... No way!

But I don't see returns as a purely random normal distribution. If the tail end edges down into negative territory because the mean shifts left, there will be plenty of resistance to keep actual returns positive. "It's the economy, stupid" is a lesson every politician has learned.
FWIW the ERN spreadsheet spits out more like 3.7-3.8% than the 3.25% I suggested. We're not arguing over a lot here. There are others here that will tell you that it needs to be 3% flat or less which doesn't make sense because 3% would last 30 years if I could just maintain the real value of my money - in TIPS for example. And surely diversifying with a little domestic and intl equities could only help over 30 years.

It doesn't make sense to assume that the odds of 1966 are the same in the future as they were historically when we know the future doesn't look as bright. Just generally speaking this requires at least a shred more caution but of course that can be hashed out mathematically as well.

Personally, I'd probably still be fine going for 4% if I knew I had the ability to reduce spending if necessary. But I'd be very sanguine about that possibility being higher than it has been historically. Somebody with 25x expenses isn't going to starve if they have some common sense. The plan needs to be reevaluated each year. Of course if this was 1985 I'd probably have been fine going with 4.5% if I had flexibility in expenses.
I agree that we're really not arguing over much. In practical terms, 3.7% and 4% (with discretionary expenses) aren't much different.
They don't have to be independent variables. They just have to not be 100% dependent. You are assuming that if the expected mean return falls from 8% to 4% that the variance is reduced exactly proportionally. That's a very strong assumption to be making without much evidence.
Last edited by skierincolorado on Wed Aug 04, 2021 5:29 pm, edited 1 time in total.
skierincolorado
Posts: 2377
Joined: Sat Mar 21, 2020 10:56 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by skierincolorado »

HomerJ wrote: Wed Aug 04, 2021 3:25 pm
alex_686 wrote: Wed Aug 04, 2021 2:55 pm
marcopolo wrote: Wed Aug 04, 2021 2:13 pm I don't have nearly the faith you have in these models. I think past history (how poorly most models have predicted market behavior) supports my skepticism.

I do agree with your last paragraph, there are so many different models out there, one can construct whatever answer they want, which I suspect is what most people do to support there own preconceived biases.
Can you explain why you prefer past history? So, my position: Models bad, history worse.

Historical models offer very little information. Worse, most people don't have a firm grasp on statistics or the underlying data. They come away worse from the experience.

There are "secular periods" (the technical economic term) in economics and investing. 5 to 15 years are common but there are examples far outside these boundaries. This period will have a mean reverting characteristic. Stock and bond returns, volatility, and correlations. Then at some random interval the economic and investing system will shift to a new state with new characteristics.

You have to keep your analysis to within that period. If you cross periods you are probably breaking your statistical assumption. You need to reach for more complex methods like a "Jump Diffusion Model".

You can't use analysis from a prior secular period for the current period. I mean, we are talking about different system. They are going to act differently.

However, you can validate a model using data from prior periods. Then use current data to run your model.

So, dig through the academic data and you will see that the models have preformed far better than historical analysis for long term planning.
Historical models when you are just looking at worst case are extremely useful. You are technically making a prediction, but a super conservative very broad one.

Imagine building a house near a river. You check the historical record of floods in the past 100 years.

Say the worst case flood ever (so far!) was 20 feet high, and that only happened once.

Building your house 22 feet above the river is a pretty conservative choice. The historical data is indeed useful and it doesn't require a model that accurately predicts how often floods will hit 8 feet or 12 feet or 15 feet. I don't care.

Building your house 17 feet above the river is a lot more risky, and you'll need a better model of what conditions caused 17-20 foot floods in the past, and how likely it is to happen in the next 10-20 years.

Sure, some variables may change, and maybe there will be a 23-foot flood in the future. That's always possible.

But if you come along with a model that says "hey, it looks like we're due for some higher than average floods going forward", I would say "Well, good thing I picked a spot higher than the worst flood so far".
Your flood analogy is perfect. So the worst flood so far was 20 feet you say? OK well now I tell you climate change is going to make floods 25% bigger on average. I don't tell you what's going to happen to your worst floods and leave you guessing. Maybe you're right and there's a damn just downstream that's 20 feet tall and if any floods are taller than 20 feet it just goes over the damn and the water stops rising and you're 22 foot above ground house is fine. But that's a pretty optimistic assumption unless you know there's a damn there and how tall it is. Maybe it's just not physically possible to rain enough to make floods bigger than 20 feet even if all the other floods are getting worse. Again - a very optimistic assumption. The smart thing to do with the information I just gave you about climate change is to build your house 25 feet above ground instead of the historical minimum of 22. Just like the smart thing to do would be to drop your withdrawal rate to 3.5 or 3.7% as Vanguard suggests - or have more flexibility in your spending than the traditional 4% rule requires.

And of course we don't really have to guess what will happen to the variance. Economic models can tell us that too... an generally speaking there's more variance when valuations are higher. Just like climate models tell us not only will the mean of floods/heatwaves etc. increase, but so will the standard deviation/variance.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by HomerJ »

skierincolorado wrote: Wed Aug 04, 2021 5:23 pm
HomerJ wrote: Wed Aug 04, 2021 3:25 pm
alex_686 wrote: Wed Aug 04, 2021 2:55 pm
marcopolo wrote: Wed Aug 04, 2021 2:13 pm I don't have nearly the faith you have in these models. I think past history (how poorly most models have predicted market behavior) supports my skepticism.

I do agree with your last paragraph, there are so many different models out there, one can construct whatever answer they want, which I suspect is what most people do to support there own preconceived biases.
Can you explain why you prefer past history? So, my position: Models bad, history worse.

Historical models offer very little information. Worse, most people don't have a firm grasp on statistics or the underlying data. They come away worse from the experience.

There are "secular periods" (the technical economic term) in economics and investing. 5 to 15 years are common but there are examples far outside these boundaries. This period will have a mean reverting characteristic. Stock and bond returns, volatility, and correlations. Then at some random interval the economic and investing system will shift to a new state with new characteristics.

You have to keep your analysis to within that period. If you cross periods you are probably breaking your statistical assumption. You need to reach for more complex methods like a "Jump Diffusion Model".

You can't use analysis from a prior secular period for the current period. I mean, we are talking about different system. They are going to act differently.

However, you can validate a model using data from prior periods. Then use current data to run your model.

So, dig through the academic data and you will see that the models have preformed far better than historical analysis for long term planning.
Historical models when you are just looking at worst case are extremely useful. You are technically making a prediction, but a super conservative very broad one.

Imagine building a house near a river. You check the historical record of floods in the past 100 years.

Say the worst case flood ever (so far!) was 20 feet high, and that only happened once.

Building your house 22 feet above the river is a pretty conservative choice. The historical data is indeed useful and it doesn't require a model that accurately predicts how often floods will hit 8 feet or 12 feet or 15 feet. I don't care.

Building your house 17 feet above the river is a lot more risky, and you'll need a better model of what conditions caused 17-20 foot floods in the past, and how likely it is to happen in the next 10-20 years.

Sure, some variables may change, and maybe there will be a 23-foot flood in the future. That's always possible.

But if you come along with a model that says "hey, it looks like we're due for some higher than average floods going forward", I would say "Well, good thing I picked a spot higher than the worst flood so far".
Your flood analogy is perfect. So the worst flood so far was 20 feet you say? OK well now I tell you climate change is going to make floods 25% bigger on average.
Heh, I figured someone was going to use climate change with my analogy.

But there is no obvious equivalent to climate change in the financial world. There MAY be some new variables that will cause worse returns in the future, worse than we've ever see in the past. Sure, that's possible, but you can't make that prediction as of now.

All you have is a prediction of low short-term returns from stocks and bonds.

That's not enough to break 4%. That's just telling me we're likely going to get a lot more rain than average so floods will worse than normal.

But you have no proof that they will be the worse in history.

4% is already prepared for low returns and/or bad Sequence of Returns.

Just like my 22 foot house is prepared for higher than average rain. Low returns or higher rain isn't enough to break the safety net.

You are making the case that the flood or the next 30 years has a decent chance to be the absolute WORST in U.S. history.

Just because "expected" returns are currently low.

Low isn't enough. We'd had low.. We'd had NEGATIVE for 16 years in bonds, and still 4% worked.

It's like the one time the flood crested 20 feet was when it rained 40 days in a row... It was a really bad year.

All you are predicting is that it might rain 20-25 days in a row instead of the normal 15... I think my 22 foot house is still pretty safe.

(Wow, I've really tortured that analogy into oblivion haven't I? Let's just agree to disagree - you can go ahead get the last word in though if you want) :)
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
seajay
Posts: 1656
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by seajay »

MarkRoulo wrote: Tue Aug 03, 2021 7:33 pm I will focus on 1966:

In 1966, 10-year US treasuries seem to have had a yield of about 4.5% (nominal)
In 1966, The US stock market seems to have traded at a P/E of about 18 (trailing)

Today: 10-year US treasuries yield about 1% (nominal).
Today: The US market P/E seems to be about 35 (trailing, I presume)

Why is it so questionable that today's market might behave worse than 1966 in real terms?
It's less a case of start date valuations, rather sequence of returns. 4% SWR and two years of -30% total returns and the SW value doubles relative to the ongoing portfolio value, as though starting with a 8% SWR. If you've held for a number of years beforehand and had seen reasonable gains, the impact of the two bad -30% years is lower. +43% gain following starting a 4% SWR and the ongoing effective SWR % reduces to 2.8%; -30% decline the next year and the effective ongoing SWR % is back up to 4% again; Yet another -30% decline and the effective SWR % rises to 5.7%. Which has a far better prospect of sustaining through than the above 8% SWR.

Higher valuations do not necessarily infer large declines are going to occur sooner rather than later, today's ceiling often becomes tomorrows floor. But yes, equally at some point when large declines do occur the immediately preceding valuations can be relatively high.

Today's 35 PE's could see that decline through increased E rather than declining P, perhaps as a consequence of higher inflation/prices.

Predicting relatively low current start date SWR % is paramount to suggesting large declines will occur sooner rather than later. Good luck with that, some were predicting long dated treasuries only had one way to go for years now. One day some will, by luck, be correct, and no doubt those in that position will proclaim how great they are at predictions. I suspect that many who more generally followed through with real world short bets on such predictions more often end up broke.
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by alex_686 »

HomerJ wrote: Wed Aug 04, 2021 3:25 pm Historical models when you are just looking at worst case are extremely useful. You are technically making a prediction, but a super conservative very broad one.
Could you please walk me through how you would build a AA and rebalancing plan based on this. I sort of understand the concept - I do work in risk management. However I don't think it is practical. But I am open to have my mind changed.

2 notes on this.

First, you want to build a plan that maximizes your chance of meeting your goals (returns) verse risk. You can't just focus on one, but rather both. The rule of thumb is that you can fund your retirement at replacement income by saving 10% to 20% gross income using conservative assumptions. Something like real long term equity returns around 5% to 7%, 60/40 portfolio, 25x portfolio level to retirement spending. If you want to be super conservative you need 40% to 60% income. Now we are assuming a bond heavy portfolio or 50x portfolio level. There is a cost to being risk adverse.

Second, historical data on crashes are bear markets is pretty thin. It depends on which data you want to included and how you want to slice it. You can either take a expansive view (past 200 to 300 years, worldwide) or narrow (modern period, past 100 years, US only). However, bull markets are similar. Crashes and bear markets are unique. The length and depth of crashes are different. Recovery periods are different. Assets work differently. Sometimes bonds are a safe asset, sometimes not. Sometimes inflation becomes a issue, other times it is deflation.

In your analogy the highest water level in the past 100 years was 20 ft. What do you think at 22 feet is conservative? How did you create your error bars. How confident are you? I lived near a town that had 4 100 year floods in under 20 years. You only have 100 spring time flood sampling periods. So small sample set. You are looking at a dynamic system. Yes, climate change. But also there have probably been changes to the watershed upstream.

Specifically, we have fact cases where stocks had negative real returns over 10 years, bonds have had negative real returns over 10 years, and both stocks and bonds have had no return for 20 years. So how is this actionable?

I have worked various risk desks. We always set up scenarios assuming the worst case scenario. In every crisis that I have worked (and that I know off) some aspect of the crash was far worse then our worse case scenario.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by HomerJ »

alex_686 wrote: Thu Aug 05, 2021 12:08 pm
HomerJ wrote: Wed Aug 04, 2021 3:25 pm Historical models when you are just looking at worst case are extremely useful. You are technically making a prediction, but a super conservative very broad one.
Could you please walk me through how you would build a AA and rebalancing plan based on this. I sort of understand the concept - I do work in risk management. However I don't think it is practical. But I am open to have my mind changed.
It's really simple. You assume low historical returns and save a lot. You then adjust on actual returns. "Expected" returns are never referenced ever.

AA is chosen on risk management, NOT expected returns.

Look, when one is young, there is no exact plan. At 35 with two young kids, your career is still progressing. You have very little idea what your retirement expenses will be, you don't know how your career is going to plateau or how much you'll be making in the future.

There is no "building a plan" around "expected" short-term returns.

You live below your means, and try to save a good chunk, and keep most of it in stocks with a bit in safer assets in case you lose your job or something.

AA can just be chosen on how close you are to retiring. 90/10 when young, moving gradually to 60/40 (or 40/60) as you approach retirement.

You get what you get, and you adjust to actual returns.

Maybe by 45-50, you have an idea of what you will need in retirement. And you can see how much you've saved.

And then you assume low returns going forward, and adjust your savings and how long you still need to work based on that.

And then again, adjust further to ACTUAL returns.

You don't save less if valuations are low and "expected" returns are high. You save always assuming low returns, and if good returns show up, you adjust AFTER that. Maybe you retire earlier or you can spend more.

Making plans on "expected" returns is foolish because the error bars are too broad. Swedroe suggested about 7-8 years back that people could invest less in stocks, if they instead over-weighted small-value, because the "expected" returns were higher. That didn't work out, because the large spread in error bars showed up. People making a plan that counted on making more ended up making less.

Just always assume low returns is the safe bet. Just like we plan around low returns for pulling money in retirement (4% or 3% whatever), use low returns for planning in accumulation as well. That's the conservative safe bet.

I don't believe it's possible to make accurate predictions that are useful for planning.

Age in bonds, you get what you get, and you adjust to actual returns is a rough investing plan, but it does work. Calculating "expected" returns is not required. Changing one's AA based on "expected" returns is a silly exercise. Make changes based on actual returns, not "expected" returns (with plus/minus 6% error bars)
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
dboeger1
Posts: 1411
Joined: Fri Jan 13, 2017 6:32 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by dboeger1 »

I realize this is more a discussion of SWR and not the nuances of how to retire at 40, but I think as the target retirement age gets earlier and life expectancy gets longer, the concept of perpetual retirement with no future earnings becomes less relevant and practical. If someone said they expected to FIRE at 20 and live to be 120, I sure hope they would plan on generating income in their 100 year retirement, unless they're really just that loaded and unmotivated. To be fair, nobody in this topic explicitly advised against earning money in retirement, but for people on the fence about the whole FIRE concept over the difference between 3% and 4% SWR, I think it's important to remember that flexible spending and additional income go a long way to improving robustness. Being able to earn even an extra $10k-$20k annually doing part time side work while cutting a few thousand in expenses can go a long way without necessarily destroying one's FIRE aspirations. Another thing to consider is that many Bogleheads in particular realistically have more backup options than they realize. There's a common idea here to not include home equity in net worth or retirement portfolio calculations, but if a reverse mortgage or downsizing and relocating to a cheaper place can keep us from living on the street and eating cat food in our old age, very few of us would refuse to do it in the end. Basically, while I appreciate SWR discussions as a starting point or guideline, I think they're a bit overly simplistic in practice and most individuals capable of retiring in any capacity by 40 will be enterprising enough to make it work in the end. Being FI earlier is ultimately better than not anyway. It allows people to take bigger career risks and potentially earn more money as a result. I would not let these kinds of findings discourage me from pursuing FIRE.
marcopolo
Posts: 8445
Joined: Sat Dec 03, 2016 9:22 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by marcopolo »

alex_686 wrote: Thu Aug 05, 2021 12:08 pm
HomerJ wrote: Wed Aug 04, 2021 3:25 pm Historical models when you are just looking at worst case are extremely useful. You are technically making a prediction, but a super conservative very broad one.
Could you please walk me through how you would build a AA and rebalancing plan based on this. I sort of understand the concept - I do work in risk management. However I don't think it is practical. But I am open to have my mind changed.

2 notes on this.

First, you want to build a plan that maximizes your chance of meeting your goals (returns) verse risk. You can't just focus on one, but rather both. The rule of thumb is that you can fund your retirement at replacement income by saving 10% to 20% gross income using conservative assumptions. Something like real long term equity returns around 5% to 7%, 60/40 portfolio, 25x portfolio level to retirement spending. If you want to be super conservative you need 40% to 60% income. Now we are assuming a bond heavy portfolio or 50x portfolio level. There is a cost to being risk adverse.

Second, historical data on crashes are bear markets is pretty thin. It depends on which data you want to included and how you want to slice it. You can either take a expansive view (past 200 to 300 years, worldwide) or narrow (modern period, past 100 years, US only). However, bull markets are similar. Crashes and bear markets are unique. The length and depth of crashes are different. Recovery periods are different. Assets work differently. Sometimes bonds are a safe asset, sometimes not. Sometimes inflation becomes a issue, other times it is deflation.

In your analogy the highest water level in the past 100 years was 20 ft. What do you think at 22 feet is conservative? How did you create your error bars. How confident are you? I lived near a town that had 4 100 year floods in under 20 years. You only have 100 spring time flood sampling periods. So small sample set. You are looking at a dynamic system. Yes, climate change. But also there have probably been changes to the watershed upstream.

Specifically, we have fact cases where stocks had negative real returns over 10 years, bonds have had negative real returns over 10 years, and both stocks and bonds have had no return for 20 years. So how is this actionable?

I have worked various risk desks. We always set up scenarios assuming the worst case scenario. In every crisis that I have worked (and that I know off) some aspect of the crash was far worse then our worse case scenario.

This "building a plan based on expected returns" sounds great in theory. But, in practice, I have only seen it used to keep telling people to work longer, save more, spend less in an ever spiraling process.

Can you point me to an example where you or any of the proponents of this idea told people in the depths of say the 2008/2009 GFC "Hey the markets just crashed and who knows where the bottom might me, that is great news because you can now safely spend more and not worry about saving as much because these low valuations mean we are going to have some great returns ahead of us".

In fact, that is actually what happened, and that advice would not have been unreasonable, but I just dont recall anyone making that case. Instead, what I recall was a lot more of the "now is the time buckle down, cut expenses and save more".

I learned in my old information theory class that if the outcome of a certain observation metric results in the same answer regardless of the value of the metric, than the observation metric does not provide any information. Since the answer seems to always be "save more", i conclude that the valuation metric is largely information free.
Once in a while you get shown the light, in the strangest of places if you look at it right.
EddyB
Posts: 2431
Joined: Fri May 24, 2013 3:43 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by EddyB »

dboeger1 wrote: Thu Aug 05, 2021 2:38 pm I realize this is more a discussion of SWR and not the nuances of how to retire at 40, but I think as the target retirement age gets earlier and life expectancy gets longer, the concept of perpetual retirement with no future earnings becomes less relevant and practical.
And if past is a useful guide, mechanical observance of very low withdrawal rates would often produce vast fortunes over that 100 years of growth!
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by alex_686 »

HomerJ wrote: Thu Aug 05, 2021 1:23 pm It's really simple. You assume low historical returns and save a lot. You then adjust on actual returns. "Expected" returns are never referenced ever.

AA is chosen on risk management, NOT expected returns.
...
AA can just be chosen on how close you are to retiring. 90/10 when young, moving gradually to 60/40 (or 40/60) as you approach retirement.

Just always assume low returns is the safe bet. Just like we plan around low returns for pulling money in retirement (4% or 3% whatever), use low returns for planning in accumulation as well. That's the conservative safe bet.nning.

Age in bonds, you get what you get, and you adjust to actual returns is a rough investing plan, but it does work. Calculating "expected" returns is not required. Changing one's AA based on "expected" returns is a silly exercise. Make changes based on actual returns, not "expected" returns (with plus/minus 6% error bars)
Let me point out 2 things.

First, you are focusing on risk, ignoring returns.

If we extend that out logically it implies that we should be 100% government bonds. After all, government bonds are safer than stocks, and we only care about risk?

Expect you suggest that you should be 90/10 when young and 60/40 when older. Why would the AA change by age? As of 2021 is a 30 year treasury safer if held by a 25 year old or a 50 year old? I don't think so.

The reasonable conclusion is that you have multiple market expectations.
1. Equities have a higher expected return than bonds.
2. Equities have a higher expected risk than bonds.
3. There are multiple paths the future might take. Being unsure you don't go 100% equities or 100% bond. Instead you chose some optimal mixture.

Now, these market expectations may be vague and unarticulated and applied in a heuristic fashion, but the above are market expectations.

Second, I suspect that your market expectations are rosier than you project.

Historically we have had 30 year periods with negative real bond returns and negative real stock returns (dividends reinvested). If you are working off the worst case historical situations then the logical answer is not to save. Once again I think you are looking at the data and heuristically generating market expectations.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
JBTX
Posts: 11227
Joined: Wed Jul 26, 2017 12:46 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by JBTX »

FiveK wrote: Tue Aug 03, 2021 1:00 pm It's all about the assumptions.

As the article states (emphasis added)
At a 30-year horizon, there is an 18% chance of running out of money. While low, this is higher than found in the original research, primarily because Vanguard assumed that stocks’ and bonds’ returns in the future would be lower than their historical averages.
While nobody can predict the future, this seems like a reasonable and conservative assumption. With starting points at near record high PES and near record low interest rates, it would seem improbable that outcomes would be distributed similarly to all historical outcomes that included many lower PE and higher interest rate starting points.
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by nigel_ht »

alex_686 wrote: Fri Aug 06, 2021 11:01 am
HomerJ wrote: Thu Aug 05, 2021 1:23 pm It's really simple. You assume low historical returns and save a lot. You then adjust on actual returns. "Expected" returns are never referenced ever.

AA is chosen on risk management, NOT expected returns.
...
AA can just be chosen on how close you are to retiring. 90/10 when young, moving gradually to 60/40 (or 40/60) as you approach retirement.
Let me point out 2 things.

First, you are focusing on risk, ignoring returns.

If we extend that out logically it implies that we should be 100% government bonds. After all, government bonds are safer than stocks, and we only care about risk?
Nobody is ignoring returns which is why some folks are willing to accept more risk than they would otherwise do in their AA. Inflation is a risk as well which is why most sane AA have a stock component.
Expect you suggest that you should be 90/10 when young and 60/40 when older. Why would the AA change by age? As of 2021 is a 30 year treasury safer if held by a 25 year old or a 50 year old? I don't think so.
Age doesn’t matter as much as time to retirement. If the 50 year old intends to retire at age 60 and the 25 year old by 35 their risk posture should be similar.
Second, I suspect that your market expectations are rosier than you project.

Historically we have had 30 year periods with negative real bond returns and negative real stock returns (dividends reinvested). If you are working off the worst case historical situations then the logical answer is not to save. Once again I think you are looking at the data and heuristically generating market expectations.
This is a bizarre conclusion. When making risk assessment many folks will look at expected case and worst case. The military plans that way.

“The activity to determine possible threat courses of action (COAs), describe threat COAs, rank COAs in probable order of adoption, and at a minimum identify the most probable and the most dangerous COAs. (FM 2-01.3) (USAICoE)”

EMLCOA (enemy most likely course of action) is what you plan against. Stocks.

EMDCOA (enemy most dangerous COA) you create contingencies for. Bonds/Gold/Cash

You hedge against the worst case but still plan for the expected case which isn’t negative returns for 30 years.
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by alex_686 »

marcopolo wrote: Thu Aug 05, 2021 3:09 pm This "building a plan based on expected returns" sounds great in theory. But, in practice, I have only seen it used to keep telling people to work longer, save more, spend less in an ever spiraling process.

Can you point me to an example where you or any of the proponents of this idea told people in the depths of say the 2008/2009 GFC "Hey the markets just crashed and who knows where the bottom might me, that is great news because you can now safely spend more and not worry about saving as much because these low valuations mean we are going to have some great returns ahead of us".

In fact, that is actually what happened, and that advice would not have been unreasonable, but I just dont recall anyone making that case. Instead, what I recall was a lot more of the "now is the time buckle down, cut expenses and save more".
I have highlighted the part that is confusing me.

Your goal is to have 200k in 10 years. You have 100k today. The market dips to 80k. How will that increase the likelihood of meeting your 200k goal in 10 years?

I am going to guess you are using some type of mean reverse argument, as it is the most common. This is not supported by either theory or historical data. The market's mean if time varying. You have one secular period with one set of returns, then another with a different set.

In-between, like the GFC, it is hard to tell what the new norm will be. The system is in flux.

Or maybe you have a recently cognitive bias. Yes, the lower ratios after the GFC indicated higher returns - in hindsight. There are other fact cases were the lower returns indicated another crash - which were obvious in hindsight.
marcopolo wrote: Thu Aug 05, 2021 3:09 pm I learned in my old information theory class that if the outcome of a certain observation metric results in the same answer regardless of the value of the metric, than the observation metric does not provide any information. Since the answer seems to always be "save more", i conclude that the valuation metric is largely information free.
Now, my day job is adjacent to this. I can tell you we are getting different answers today then we were getting 2 years ago.

I think the answer you are looking for lays in psychology and risk aversion. People have a tendency to crank up their goals. People will set their goals for 1m and work really hard towards that goal. A few good years and now 2m seem attenable, so they move the goal posts to 2m. That is the new minimum goal.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by HomerJ »

alex_686 wrote: Fri Aug 06, 2021 11:01 am
HomerJ wrote: Thu Aug 05, 2021 1:23 pm It's really simple. You assume low historical returns and save a lot. You then adjust on actual returns. "Expected" returns are never referenced ever.

AA is chosen on risk management, NOT expected returns.
...
AA can just be chosen on how close you are to retiring. 90/10 when young, moving gradually to 60/40 (or 40/60) as you approach retirement.

Just always assume low returns is the safe bet. Just like we plan around low returns for pulling money in retirement (4% or 3% whatever), use low returns for planning in accumulation as well. That's the conservative safe bet.nning.

Age in bonds, you get what you get, and you adjust to actual returns is a rough investing plan, but it does work. Calculating "expected" returns is not required. Changing one's AA based on "expected" returns is a silly exercise. Make changes based on actual returns, not "expected" returns (with plus/minus 6% error bars)
Let me point out 2 things.

First, you are focusing on risk, ignoring returns.

If we extend that out logically it implies that we should be 100% government bonds. After all, government bonds are safer than stocks, and we only care about risk?

Expect you suggest that you should be 90/10 when young and 60/40 when older. Why would the AA change by age? As of 2021 is a 30 year treasury safer if held by a 25 year old or a 50 year old? I don't think so.
Because I'm talking about short-term risk.

It's really not that hard... When I say pick an AA based on risk, I mean personal risk, which changes as you age and get closer to retirement.

Pick an AA assuming the market might crash 50% (or more) tomorrow and take 5-10 or more years to recover. Or trend sideways and lose to inflation for 5-10 or more years.

Because it might.

This is always true, regardless of valuations or any other metrics.

If you're young, you can be 90% stocks in your retirement accounts, even assuming a crash is coming tomorrow because you don't need to touch that money for years and years. So you're not as worried about short-term risks.

Keep SOME in safer assets, for an emergency fund, or if you lose your job, but most can be long-term in stocks.

As you get closer to retirement, you move more to safer assets, because if you retire and there's a crash, you'll need money to live on while you wait for stocks to recover. Or even if you're not retired, but older, you should have more in safer assets, because if you lose your job in a stock crash while older, it may be harder to find another job. You may end up effectively retired before you plan to.

All these risks are age-related, and have nothing to do with current market valuations.

You may not agree, but this shouldn't be hard to understand.
The reasonable conclusion is that you have multiple market expectations.
1. Equities have a higher expected return than bonds.
2. Equities have a higher expected risk than bonds.
3. There are multiple paths the future might take. Being unsure you don't go 100% equities or 100% bond. Instead you chose some optimal mixture.

Now, these market expectations may be vague and unarticulated and applied in a heuristic fashion, but the above are market expectations.
My market assumptions are:

Long-term equity returns will be positive real.
Equities will do better than bonds long-term

That's about it.

I pick an AA based on my personal risk situation, assuming a bad market for the next 10+ years starts tomorrow. I try to keep as much as possible in equities, because I assume long-term they will do better than bonds, while keeping enough in bonds to protect me from a short-to-medium bad period in stocks.

And then I don't have to worry about predicting a bad market, because I'm already prepared for it.

Second, I suspect that your market expectations are rosier than you project.

Historically we have had 30 year periods with negative real bond returns and negative real stock returns (dividends reinvested). If you are working off the worst case historical situations then the logical answer is not to save. Once again I think you are looking at the data and heuristically generating market expectations.
Which 30-year period did we have negative real stock market returns with dividends reinvested?
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by HomerJ »

nigel_ht wrote: Fri Aug 06, 2021 11:27 amAge doesn’t matter as much as time to retirement. If the 50 year old intends to retire at age 60 and the 25 year old by 35 their risk posture should be similar.
This is correct. I was using age as a proxy to how close one is to retirement.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
reln
Posts: 718
Joined: Fri Apr 19, 2019 4:01 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by reln »

Grifin wrote: Tue Aug 03, 2021 12:43 pm This morning I read a provoking article about some standard Boglehead fare; SWR and 30+ year retirement timelines. According to a study conducted by Vanguard, the risk of running out of money increases greatly with a 4% SWR from a 30 to 50 year time period assuming lower future returns which the firm expects.

Conceptually, I can relate to the thesis however, I was taken aback by the magnitude of failure rates that the study concluded for the 30, 40 & 50yr time periods

According to the article, the chance of running out of money in 30 years is 18%. "At the 40-year horizon, the failure rate rises to 46%. And at the 50-year horizon, it is 64%."

link:
https://www.marketwatch.com/story/the-f ... retirement

These figures differ greatly from many popular retirement modelling tools, ie. FireCalc. In true MW fashion, there is no real detail offered to understand the methods, data and assumptions for the analysis.

I thought this article to gin up some good banter on the board.

thoughts?

Cheers!
This is the never ending topic of BHs. To 4% or not to 4%.
reln
Posts: 718
Joined: Fri Apr 19, 2019 4:01 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by reln »

HomerJ wrote: Fri Aug 06, 2021 12:26 pm
nigel_ht wrote: Fri Aug 06, 2021 11:27 amAge doesn’t matter as much as time to retirement. If the 50 year old intends to retire at age 60 and the 25 year old by 35 their risk posture should be similar.
This is correct. I was using age as a proxy to how close one is to retirement.
The number of years you'll have in retirement matter too. A 65 year old recently retired person can expect to live say 25-30 years in retirement whereas as 35 year old can have 55-60 years in retirement. Their AA should be different because their time horizons are very different. Generally, the longer your horizon, the more stocks. So a FIRE person should have more stocks to fund a long retirement. Likewise an older retiree that has a large legacy goal effectively has a longer duration than his or her life expectancy so they should hold more stocks in their legacy portfolio bringing their overall AA higher into stocks.

No need to rebuttal. I know y'all don't agree. Just chiming in while I listen to music.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by HomerJ »

reln wrote: Fri Aug 06, 2021 12:46 pm
HomerJ wrote: Fri Aug 06, 2021 12:26 pm
nigel_ht wrote: Fri Aug 06, 2021 11:27 amAge doesn’t matter as much as time to retirement. If the 50 year old intends to retire at age 60 and the 25 year old by 35 their risk posture should be similar.
This is correct. I was using age as a proxy to how close one is to retirement.
The number of years you'll have in retirement matter too. A 65 year old recently retired person can expect to live say 25-30 years in retirement whereas as 35 year old can have 55-60 years in retirement. Their AA should be different because their time horizons are very different. Generally, the longer your horizon, the more stocks. So a FIRE person should have more stocks to fund a long retirement. Likewise an older retiree that has a large legacy goal effectively has a longer duration than his or her life expectancy so they should hold more stocks in their legacy portfolio bringing their overall AA higher into stocks.

No need to rebuttal. I know y'all don't agree. Just chiming in while I listen to music.
Heh, I agree with you. It's more complicated the younger one retires.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
randomguy
Posts: 11295
Joined: Wed Sep 17, 2014 9:00 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by randomguy »

skierincolorado wrote: Wed Aug 04, 2021 1:36 pm

And the Vanguard, nor the ERN series, are not based on the mean expectation being 1929 either. The mean expectation is only somewhat lower than in the past - and with very strong evidence that expectations should be reduced. If the mean is reduced it is just basic common sense that the worst cases and the entire distribution of outcomes is reduced as well. Unless you have a strong argument that the variance is reduced as well. But I've heard no such argument other than a guess that the U.S. govt might do something. But the U.S. govt couldn't/didn't stop 1929 or 1966 did they?
The mean expectation is 50% lower than the historical number. That is pretty close to 1929:) 4% nominal for 10 years is something that hasn't shown up very often in US history. Pretty sure it is 2000-9 and a couple of the time periods including 1930,1931. The US government didn't stop 1929 or 1966 (and may have caused both of them). They may have stopped 2000 and 2008 though from spiraling out of control.

With all these models the question is about the inputs. What if the mean is actually 6%(i.e. still a 40% reduction) instead of 4%? What if the volatility is 10% less than they are assuming. When dealing with tail end of graphs, you can get some pretty major changes with pretty small changes to the inputs especially when you amplify the results over 50 years...
seajay
Posts: 1656
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by seajay »

randomguy wrote: Fri Aug 06, 2021 3:37 pmThe mean expectation is 50% lower than the historical number.
Or that cost/taxes options are lower nowadays. Historically investing was relatively expensive, very wide market makers spreads, high brokers fees, less tax efficient options available. 4% historic real SWR is based on a real gross figure.

Jack Bogle used a historic example of how investors took on 100% of the risk for 40% of the reward, 50 year average investment lifetime, 8% gross versus 6% net rewards, 47 versus 18 gain factors. In practice many investors further lagged the 'average' due to emotional decisions, buying high, capitulating low, many ended up worse off than if they'd solely stuck to cash deposit investments.

Uncoincidentally high inflation/high yields tends to correlate with higher taxation.
User avatar
Rowan Oak
Posts: 851
Joined: Mon May 09, 2016 2:11 pm
Location: Yoknapatawpha

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by Rowan Oak »

HomerJ wrote: Wed Aug 04, 2021 3:25 pm Historical models when you are just looking at worst case are extremely useful. You are technically making a prediction, but a super conservative very broad one.

Imagine building a house near a river. You check the historical record of floods in the past 100 years.

Say the worst case flood ever (so far!) was 20 feet high, and that only happened once.

Building your house 22 feet above the river is a pretty conservative choice. The historical data is indeed useful and it doesn't require a model that accurately predicts how often floods will hit 8 feet or 12 feet or 15 feet. I don't care.

Building your house 17 feet above the river is a lot more risky, and you'll need a better model of what conditions caused 17-20 foot floods in the past, and how likely it is to happen in the next 10-20 years.

Sure, some variables may change, and maybe there will be a 23-foot flood in the future. That's always possible.

But if you come along with a model that says "hey, it looks like we're due for some higher than average floods going forward", I would say "Well, good thing I picked a spot higher than the worst flood so far".
Good analogy. Going to steal this one if you don't mind. :sharebeer
“If you can get good at destroying your own wrong ideas, that is a great gift.” – Charlie Munger
LookingForward
Posts: 58
Joined: Fri Jul 01, 2016 9:27 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by LookingForward »

alex_686 wrote: Wed Aug 04, 2021 1:57 pm
LookingForward wrote: Tue Aug 03, 2021 8:42 pm
alex_686 wrote: Tue Aug 03, 2021 12:57 pm One can make relatively decent 10 year forecasts.
Is this true?

Can one predict the most recent 10 years from the previous data? And so on?

I’ve been seeing dire predictions for the market and the economy, supposedly based on hard data, since I started paying attention, but that was only a dozen years ago.
Yes. But not on the past 10 year's worth of data. More on that latter.
I wasn't asking to use data in 10-year chunks. I was saying "could you have correctly predicted the market's performance for 2011-2021, from previous data?"

I'm pretty sure the answer is "no", and it sounds like the answer is "no, but now we have all the answers for the future". I am skeptical.
alex_686 wrote: Wed Aug 04, 2021 1:57 pm However, this is complex, nuanced, requires some subjective decisions, and requires proprietary data.
"Requires proprietary data" makes my BS meter peg at the top.
McDougal
Posts: 557
Joined: Tue Feb 27, 2018 2:42 pm
Location: Atlanta

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by McDougal »

nigel_ht wrote: Tue Aug 03, 2021 2:06 pm
young-ish wrote: Tue Aug 03, 2021 2:02 pm
skierincolorado wrote: Tue Aug 03, 2021 1:39 pm
Grifin wrote: Tue Aug 03, 2021 12:43 pm This morning I read a provoking article about some standard Boglehead fare; SWR and 30+ year retirement timelines. According to a study conducted by Vanguard, the risk of running out of money increases greatly with a 4% SWR from a 30 to 50 year time period assuming lower future returns which the firm expects.

Conceptually, I can relate to the thesis however, I was taken aback by the magnitude of failure rates that the study concluded for the 30, 40 & 50yr time periods

According to the article, the chance of running out of money in 30 years is 18%. "At the 40-year horizon, the failure rate rises to 46%. And at the 50-year horizon, it is 64%."

link:
https://www.marketwatch.com/story/the-f ... retirement

These figures differ greatly from many popular retirement modelling tools, ie. FireCalc. In true MW fashion, there is no real detail offered to understand the methods, data and assumptions for the analysis.

I thought this article to gin up some good banter on the board.

thoughts?

Cheers!
It's pretty well established that 4% is not a good withdrawal rate for a retirement over 30 years anymore. For 30 years I think 3.25% makes sense, more if SS will be significant to you. For 60 years I think 3% is necessary if you are completely giving up all work forever and have zero flexibiltiy, or 3.25% if you can have some flexibility in your spending or pick up a side gig if the market drops in the first 15 years.
How did you arrive at 3.25%? Is that a constant WD % or a starting with inflation adjusted increases? What asset allocation supports this WDR?
Image

https://earlyretirementnow.com/2016/12/ ... t-1-intro/
I have always found this chart to be an interesting way to guesstimate a SWR, or even begin to think about planning the decumulation phase. Not the only factor, but one of many. I do wonder if including the past 5+ years of market activity would alter the chart in any way.
59Gibson
Posts: 1386
Joined: Mon Dec 07, 2020 7:41 am

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by 59Gibson »

dboeger1 wrote: Thu Aug 05, 2021 2:38 pm I realize this is more a discussion of SWR and not the nuances of how to retire at 40, but I think as the target retirement age gets earlier and life expectancy gets longer, the concept of perpetual retirement with no future earnings becomes less relevant and practical. If someone said they expected to FIRE at 20 and live to be 120, I sure hope they would plan on generating income in their 100 year retirement, unless they're really just that loaded and unmotivated. To be fair, nobody in this topic explicitly advised against earning money in retirement, but for people on the fence about the whole FIRE concept over the difference between 3% and 4% SWR, I think it's important to remember that flexible spending and additional income go a long way to improving robustness. Being able to earn even an extra $10k-$20k annually doing part time side work while cutting a few thousand in expenses can go a long way without necessarily destroying one's FIRE aspirations. Another thing to consider is that many Bogleheads in particular realistically have more backup options than they realize. There's a common idea here to not include home equity in net worth or retirement portfolio calculations, but if a reverse mortgage or downsizing and relocating to a cheaper place can keep us from living on the street and eating cat food in our old age, very few of us would refuse to do it in the end. Basically, while I appreciate SWR discussions as a starting point or guideline, I think they're a bit overly simplistic in practice and most individuals capable of retiring in any capacity by 40 will be enterprising enough to make it work in the end. Being FI earlier is ultimately better than not anyway. It allows people to take bigger career risks and potentially earn more money as a result. I would not let these kinds of findings discourage me from pursuing FIRE.
+1000. Very true. It's not necessarily about quitting all paid work. I think reaching FI relatively early(under 45-50) gives one options to design their life as they see fit. Maybe you cannot fully quit work forever, but you sure as hell can get out of a job you hate and only show up for the paycheck. Example would be a 40 yr old w/ $800k portfolio and 50k/yr expenses. Maybe they're in a high pressure sales/mgmt job making $120k and been saving $40k/yr. They've had enough of it...This person could very safely pull $20k from the portfolio and get a job making $35k doing something much less stressful and hopefully enjoyable.
JPM
Posts: 586
Joined: Sun Aug 19, 2018 2:29 pm

Re: MarketWatch article about FIRE, 4% SWR and 30+ yr retirement

Post by JPM »

Just looking at the 2021 RMD divisors and adding .9 per year, the simple divisor progression listed, the divisor for 70 is about 27.4, adding simple .9 per year would get to 36.4 as a divisor at 60 and 45.4 as a divisor at 50 for the FIRE crowd.
Post Reply