Watch out for hidden risk tolerance assumptions in SWR claims

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
dboeger1
Posts: 1259
Joined: Fri Jan 13, 2017 7:32 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by dboeger1 »

nigel_ht wrote: Fri Aug 05, 2022 1:03 pm This is solved by reaching FI by age 50 and waiting for the next recession and bear. If you still have your number and want to retire then that’s a good time.
It's not FI if you have to wait for the next recession/bear for your numbers to hold. The point of reducing the WR is to account for lower returns based on current high valuations if you were to retire today.
marcopolo wrote: Fri Aug 05, 2022 1:13 pm People insist that because we are likely to get poor results, one should work longer to lower their WR. And then immediately turn around and claim doing so should be a piece of cake because the good market returns will get them there in no time. You can't have it both ways.

If you are worried about 4% not working because of low returns going forward, then the alternative is not working just a few more years to get to 3% (which might be an easy trade off), but rather working a decade or longer to lower your risk somewhat (maybe not such an easy decision now).
To be fair, I did oversimplify in my original post, but I think it's important to consider that should the market fall to lower PE ratios, the returns on those purchases would likely be higher on average (again, keeping in mind that there is still a very wide distribution of outcomes). In ERN's SWR series, he actually suggests adjustments to the SWR based on CAPE bands, so in a lower CAPE band, you might not need as low as a 3% SWR or more. So the 3% wasn't a constant target, it was more of a suggestion for someone trying to reach FI in a bull market with high CAPE. It may very well be that the hypothetical person continuing to work would be stuck at 4% WR after 4 more years, but that might be enough given reduced valuations. I realize I didn't state it that way, so that's my fault.

Ultimately, the universe doesn't care how long it takes to reach 3% WR from 4%. It might take one person 2 years, and another 20 years. Either way, if 4% isn't low enough, then it isn't low enough. It's up to each person to decide the tradeoffs they want to make. If you were to ask me personally whether I'd work 20 years to drop from 4% to 3% WR, I would probably say no. I think I could most likely make 4% work, and I also have plans to make some income in early retirement through passion projects and small businesses which I have not built into my financial assumptions. But would I work 2 years to reach 3%? Most likely yes. I don't want to work another 2 years, but the difference in peace of mind would be huge. Again, the only point I was really trying to make is that 4% is likely not as safe as many people realize, and it might be worth working a little longer to reach that 3%.
marcopolo
Posts: 6469
Joined: Sat Dec 03, 2016 10:22 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by marcopolo »

dboeger1 wrote: Fri Aug 05, 2022 1:32 pm
nigel_ht wrote: Fri Aug 05, 2022 1:03 pm This is solved by reaching FI by age 50 and waiting for the next recession and bear. If you still have your number and want to retire then that’s a good time.
It's not FI if you have to wait for the next recession/bear for your numbers to hold. The point of reducing the WR is to account for lower returns based on current high valuations if you were to retire today.
marcopolo wrote: Fri Aug 05, 2022 1:13 pm People insist that because we are likely to get poor results, one should work longer to lower their WR. And then immediately turn around and claim doing so should be a piece of cake because the good market returns will get them there in no time. You can't have it both ways.

If you are worried about 4% not working because of low returns going forward, then the alternative is not working just a few more years to get to 3% (which might be an easy trade off), but rather working a decade or longer to lower your risk somewhat (maybe not such an easy decision now).
To be fair, I did oversimplify in my original post, but I think it's important to consider that should the market fall to lower PE ratios, the returns on those purchases would likely be higher on average (again, keeping in mind that there is still a very wide distribution of outcomes). In ERN's SWR series, he actually suggests adjustments to the SWR based on CAPE bands, so in a lower CAPE band, you might not need as low as a 3% SWR or more. So the 3% wasn't a constant target, it was more of a suggestion for someone trying to reach FI in a bull market with high CAPE. It may very well be that the hypothetical person continuing to work would be stuck at 4% WR after 4 more years, but that might be enough given reduced valuations. I realize I didn't state it that way, so that's my fault.

Ultimately, the universe doesn't care how long it takes to reach 3% WR from 4%. It might take one person 2 years, and another 20 years. Either way, if 4% isn't low enough, then it isn't low enough. It's up to each person to decide the tradeoffs they want to make. If you were to ask me personally whether I'd work 20 years to drop from 4% to 3% WR, I would probably say no. I think I could most likely make 4% work, and I also have plans to make some income in early retirement through passion projects and small businesses which I have not built into my financial assumptions. But would I work 2 years to reach 3%? Most likely yes. I don't want to work another 2 years, but the difference in peace of mind would be huge. Again, the only point I was really trying to make is that 4% is likely not as safe as many people realize, and it might be worth working a little longer to reach that 3%.
How does one get from 4% to 3% by "working a little longer"?
Once in a while you get shown the light, in the strangest of places if you look at it right.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

marcopolo wrote: Fri Aug 05, 2022 1:13 pm You see no irony in saying people should work longer to get to 3% because poor returns are likely when markets are near their highs, and then saying it should only take a few years to get there because of the good market returns?!?
Nope. I’m not the op but it makes sense.

Most FIRE folks need 33x vs 25x because they have a longer retirement period. So 3% was the right target all along.

So a $1m target for $40K a year spend turns into $1.33M. If the 3% SWR worked in 1929 and 1966 then for 50 years (or whatever) unless you retired into a cohort that experiences worse than 1929 and 1966 then $1.33M is safe. Hence the name.

It is certainly faster to get from $1M to $1.33M when returns are good…as long as your luck holds you can hit your new target before the next correction to retire safely.

If not and the market crashes you go backwards and have wait for the market to recover.

Or the market stagnates and you don’t get a lot of help to get to $1.33M and have to save $333K the hard way.
seajay
Posts: 729
Joined: Sat May 01, 2021 3:26 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by seajay »

dboeger1 wrote: Fri Aug 05, 2022 12:59 pmThe whole 3% SWR suggestion was not about minimizing work, it was about being safe enough for even very long retirements through historically bad conditions
British FT30 index has lagged the Dow Industrial average (also 30 stocks) since the mid 1930's. But so also did the Dow Utilities, and Dow Transport averages. Historic UK taxation/costs issues were a considerable drag factor, in postal trading days market makers might widen spreads out to 10% or more, and brokers took much heavier amounts than nowadays. The US has been a right-tail good/great case over the last century.

19th/20th centuries and with money pegged to gold there was broad 0% inflation and government bonds yielded 4% average, which was like a 4% real yield benefit. Since WW1 and that ended, where subsequently government bonds broadly yield 0% real.

If you now lend to the government, buy treasuries, you are lending to someone who can direct inflation (print/spend money that devalues all other notes in circulation), revise the taxation rates/rules, set interest rates. If the market are presently pricing bonds to 4% yields then a $100 face value bond paying 4% yield will sell for $100 and generate a $4 interest payment. When that nominal value/yield is deflated over time, perhaps a 20 year bond and 2% yearly inflation, then by the end of that 20 years you get your $100 back, along with the final $4 interest payment, but where inflation had deflated the purchase power of those by a third. And that assumes cost free values, that are typically worse after costs and taxes are factored in. Not a particularly good deal, but is supported by pension funds/pots - that are often forced to lend to the government by law, as is the case in many countries.

Referencing/measuring historic rewards, 4% real easily being achieved, especially under right-tail/best-case conditions (US), excluding costs.taxes, and projecting that into forward time - might be considered as being optimistic. A more reasonable figure might be to assume a inflation pacing target total return, looking to minimize costs/taxes, and smooth down volatility as much as possible, in which case across 30 years you might draw a 3.3% SWR, leaving nowt remaining at the end of the 30 years.

If you attach a 0% real assumption to each asset, but where that comes with individual asset relatively high volatility, then diversification can help smooth down that volatility risk. Some land, stocks, commodity, bonds and where as one falters another might pick up the slack.
marcopolo
Posts: 6469
Joined: Sat Dec 03, 2016 10:22 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by marcopolo »

nigel_ht wrote: Fri Aug 05, 2022 1:44 pm
marcopolo wrote: Fri Aug 05, 2022 1:13 pm You see no irony in saying people should work longer to get to 3% because poor returns are likely when markets are near their highs, and then saying it should only take a few years to get there because of the good market returns?!?
Nope. I’m not the op but it makes sense.

Most FIRE folks need 33x vs 25x because they have a longer retirement period. So 3% was the right target all along.

So a $1m target for $40K a year spend turns into $1.33M. If the 3% SWR worked in 1929 and 1966 then for 50 years (or whatever) unless you retired into a cohort that experiences worse than 1929 and 1966 then $1.33M is safe. Hence the name.

It is certainly faster to get from $1M to $1.33M when returns are good…as long as your luck holds you can hit your new target before the next correction to retire safely.

If not and the market crashes you go backwards and have wait for the market to recover.

Or the market stagnates and you don’t get a lot of help to get to $1.33M and have to save $333K the hard way.

You completely missed my point.

The post I was responding to was making the reasonable argument that you need 3% WR rather than 4% because of low expected forward returns.
Then in the next breath said it was easy to get to 3% in a short few years because good market returns would do most of the work for you.

My point was not against needing 3%, but rather in the trivializing what it might take to get there.
If you are expecting poor returns and want to get to a lower WR because of that, you have to consider the same low-return environment when estimating how long it will take you to get there.
Once in a while you get shown the light, in the strangest of places if you look at it right.
Northern Flicker
Posts: 11031
Joined: Fri Apr 10, 2015 12:29 am

Re: "Stocks increase SWR" assumes a risk-tolerant investor

Post by Northern Flicker »

nisiprius wrote: Thu Aug 04, 2022 7:32 am
muffins14 wrote: Thu Aug 04, 2022 7:23 am I’d recommend removing everything you have on the 7% withdrawal rate to make the post more concise. No one here is going to plan for 7%, so it’s a bit of a distraction
My point is that stocks make a huge increase in SWR for a very-risk-tolerant investor, and very little difference for a risk-averse investor. What you see is going to depend on what you are assuming for risk tolerance.
Risk tolerance is not the only issue, or maybe the notion of risk needs to be enhanced to cover variability in spending needs. You may plan on a 3.25% or 4% SWR over 40 years, but ultimately you don't have full control of your spending glide path. An illness and long-term care could mean that you end up needing a 12% SWR over 10 years.
Last edited by Northern Flicker on Fri Aug 05, 2022 2:22 pm, edited 1 time in total.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

marcopolo wrote: Fri Aug 05, 2022 1:17 pm
nigel_ht wrote: Fri Aug 05, 2022 1:03 pm
I’d rather discount my portfolio if I was feeling antsy rather than reduce my SWR because SWR is computed based on historical worst case.

That means that SWR is based on the outcome of folks that would have retired in 1929 and 1966. When we have enough data also 2008 and 2000…
What is the difference between these two?
You arrive at the same dollar number in either case for the first year. From there it is just a inflation adjustment of that dollar amount.

I am not seeing how the two would differ in any way.
I am aged 60 and my number to retire is $40K a year income from my portfolio for the next 30 years.

This is my go-no go decision criteria.

$1M discounted 25% = $750K @ 4% = $30K a year. Fail.
$1M at 3% = $30K a year. Fail.

I save another $333K.

$1.33M discounted to $1M @ 4% = $40K. Pass
$1.33M @ 3% = $40K (close enough). Pass

In actual retirement:

$1.33M @ 4% = $53,200 in withdrawals
$1.33M @ 3% = $39,900 in withdrawals

Remember that SWR is a ceiling and not a floor. It is not a required draw in execution.

Given that 4% SWR (okay, slightly less) worked for 1929 and 1966 then $50K is my safe ceiling.

If SORR doesn't strike I can happily pull $50K a year if I want because I didn't drop my SWR to 3% but instead discounted my portfolio.
If SORR strikes and I get antsy I can pull only $40K a year.

It's kinda like if you want to raise or lower the volume for a particular singer. I can do so at the input for her mic or the master for everyone. Her final volume is the same but where I choose to make the modification changes the final sound.

I could just discount the equity portion because of high valuations and leave the fixed income alone. That lowers my target from $1.33M to $1.2M...so in the sound board analogy using a discount is like changing the volume for a single input vs changing SWR which is more like messing with the master.
marcopolo
Posts: 6469
Joined: Sat Dec 03, 2016 10:22 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by marcopolo »

nigel_ht wrote: Fri Aug 05, 2022 2:17 pm
marcopolo wrote: Fri Aug 05, 2022 1:17 pm
nigel_ht wrote: Fri Aug 05, 2022 1:03 pm
I’d rather discount my portfolio if I was feeling antsy rather than reduce my SWR because SWR is computed based on historical worst case.

That means that SWR is based on the outcome of folks that would have retired in 1929 and 1966. When we have enough data also 2008 and 2000…
What is the difference between these two?
You arrive at the same dollar number in either case for the first year. From there it is just a inflation adjustment of that dollar amount.

I am not seeing how the two would differ in any way.
I am aged 60 and my number to retire is $40K a year income from my portfolio for the next 30 years.

This is my go-no go decision criteria.

$1M discounted 25% = $750K @ 4% = $30K a year. Fail.
$1M at 3% = $30K a year. Fail.

I save another $333K.

$1.33M discounted to $1M @ 4% = $40K. Pass
$1.33M @ 3% = $40K (close enough). Pass

In actual retirement:

$1.33M @ 4% = $53,200 in withdrawals
$1.33M @ 3% = $39,900 in withdrawals

Remember that SWR is a ceiling and not a floor. It is not a required draw in execution.

Given that 4% SWR (okay, slightly less) worked for 1929 and 1966 then $50K is my safe ceiling.

If SORR doesn't strike I can happily pull $50K a year if I want because I didn't drop my SWR to 3% but instead discounted my portfolio.
If SORR strikes and I get antsy I can pull only $40K a year.

It's kinda like if you want to raise or lower the volume for a particular singer. I can do so at the input for her mic or the master for everyone. Her final volume is the same but where I choose to make the modification changes the final sound.

I could just discount the equity portion because of high valuations and leave the fixed income alone. That lowers my target from $1.33M to $1.2M...so in the sound board analogy using a discount is like changing the volume for a single input vs changing SWR which is more like messing with the master.
You have $1.33M either way.
How "safe" you are is going to be the same regardless of how you choose to view your withdrawals.
Yes. Withdrawing less is safer than withdrawing more. That is true whether you use lower SWR or discount your portfolio.
Look up "Associative property of multiplication"
Once in a while you get shown the light, in the strangest of places if you look at it right.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

marcopolo wrote: Fri Aug 05, 2022 2:02 pm
nigel_ht wrote: Fri Aug 05, 2022 1:44 pm
marcopolo wrote: Fri Aug 05, 2022 1:13 pm You see no irony in saying people should work longer to get to 3% because poor returns are likely when markets are near their highs, and then saying it should only take a few years to get there because of the good market returns?!?
It is certainly faster to get from $1M to $1.33M when returns are good…as long as your luck holds you can hit your new target before the next correction to retire safely.

If not and the market crashes you go backwards and have wait for the market to recover.

Or the market stagnates and you don’t get a lot of help to get to $1.33M and have to save $333K the hard way.
You completely missed my point.
Did I?
The post I was responding to was making the reasonable argument that you need 3% WR rather than 4% because of low expected forward returns.
Then in the next breath said it was easy to get to 3% in a short few years because good market returns would do most of the work for you.
This would have been a reasonable statement in 2021. In 2022 it's a little late for that.

If you had $1M in Jan 2021 you might consider retiring but the valuations bothered you. Rather than retiring right then you decided OMY and shoot for $1.33M while saving $50K a year. If 2021 will be like the last half of 2020 it shouldn't take long right?

If you had retired in December 31, 2021 you would use a 4% rate: $1M - $40K = $960K * 1.28 (28% S&P 500 gains) = $1,228,800
Since you worked one more year you have $1M * 1.28 = 1.28M + $50K = $1.33M.

You retire at a 3% rate. Win!

Is it somewhat contrived? Only the $50K savings part...but most folks shooting for FIRE have fairly high savings rates and this actually could have happened last year.

Both scenarios probably work out fine because it is rare to get worse than historical worst case. 2022 might be worse than 1966 but its still not very likely.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

marcopolo wrote: Fri Aug 05, 2022 2:27 pm
nigel_ht wrote: Fri Aug 05, 2022 2:17 pm
marcopolo wrote: Fri Aug 05, 2022 1:17 pm
nigel_ht wrote: Fri Aug 05, 2022 1:03 pm
I’d rather discount my portfolio if I was feeling antsy rather than reduce my SWR because SWR is computed based on historical worst case.

That means that SWR is based on the outcome of folks that would have retired in 1929 and 1966. When we have enough data also 2008 and 2000…
What is the difference between these two?
You arrive at the same dollar number in either case for the first year. From there it is just a inflation adjustment of that dollar amount.

I am not seeing how the two would differ in any way.
I am aged 60 and my number to retire is $40K a year income from my portfolio for the next 30 years.

This is my go-no go decision criteria.

$1M discounted 25% = $750K @ 4% = $30K a year. Fail.
$1M at 3% = $30K a year. Fail.

I save another $333K.

$1.33M discounted to $1M @ 4% = $40K. Pass
$1.33M @ 3% = $40K (close enough). Pass

In actual retirement:

$1.33M @ 4% = $53,200 in withdrawals
$1.33M @ 3% = $39,900 in withdrawals

Remember that SWR is a ceiling and not a floor. It is not a required draw in execution.

Given that 4% SWR (okay, slightly less) worked for 1929 and 1966 then $50K is my safe ceiling.

If SORR doesn't strike I can happily pull $50K a year if I want because I didn't drop my SWR to 3% but instead discounted my portfolio.
If SORR strikes and I get antsy I can pull only $40K a year.

It's kinda like if you want to raise or lower the volume for a particular singer. I can do so at the input for her mic or the master for everyone. Her final volume is the same but where I choose to make the modification changes the final sound.

I could just discount the equity portion because of high valuations and leave the fixed income alone. That lowers my target from $1.33M to $1.2M...so in the sound board analogy using a discount is like changing the volume for a single input vs changing SWR which is more like messing with the master.
You have $1.33M either way.
How "safe" you are is going to be the same regardless of how you choose to view your withdrawals.
Yes. Withdrawing less is safer than withdrawing more. That is true whether you use lower SWR or discount your portfolio.
Look up "Associative property of multiplication"
Nope. You're ignoring that the discount is used in the decision stage and not the withdrawal phase.

You are also ignoring that the discount can only be applied to stocks and not bonds.

So for a 60/40 portfolio discounting because of high valuations only applies to 60% of the portfolio value and the new target is $1.2M rather than $1.33M.

So clearly you do not have "1.33M either way".
saver7007
Posts: 40
Joined: Thu Jan 06, 2022 2:30 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by saver7007 »

nisiprius wrote: Fri Aug 05, 2022 6:35 am In other words, the required planning time frame for an individual is not-quite-double average life expectancy.
That would make a pretty awesome rule of thumb for (early) retirement planning, I just went and looked this up for myself.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

dboeger1 wrote: Fri Aug 05, 2022 1:32 pm
nigel_ht wrote: Fri Aug 05, 2022 1:03 pm This is solved by reaching FI by age 50 and waiting for the next recession and bear. If you still have your number and want to retire then that’s a good time.
It's not FI if you have to wait for the next recession/bear for your numbers to hold. The point of reducing the WR is to account for lower returns based on current high valuations if you were to retire today.
Sure you are. Historically you are FI.

However, even though it worked in 1966 and 1929, retiring now doesn't allow you to SWAN.

Waiting for the next recession/bear and then retiring if you still have your target number is more psychological than financial unless you really wanted to guard against worse than historical outcomes.

This is no different than waiting until you have 3% (for a 30 year retirement) even though 4% (slightly less) works.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

saver7007 wrote: Fri Aug 05, 2022 2:48 pm
nisiprius wrote: Fri Aug 05, 2022 6:35 am In other words, the required planning time frame for an individual is not-quite-double average life expectancy.
That would make a pretty awesome rule of thumb for (early) retirement planning, I just went and looked this up for myself.
Well...maybe not...

https://www.ssa.gov/oact/STATS/table4c6.html

If you're 20 your life expectancy is another 57 years. I think 114 years might be a bit ambitious.

Even for a 40 yo FIRE retiree... 77 years is a bit long...
dboeger1
Posts: 1259
Joined: Fri Jan 13, 2017 7:32 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by dboeger1 »

marcopolo wrote: Fri Aug 05, 2022 2:02 pm
nigel_ht wrote: Fri Aug 05, 2022 1:44 pm
marcopolo wrote: Fri Aug 05, 2022 1:13 pm You see no irony in saying people should work longer to get to 3% because poor returns are likely when markets are near their highs, and then saying it should only take a few years to get there because of the good market returns?!?
Nope. I’m not the op but it makes sense.

Most FIRE folks need 33x vs 25x because they have a longer retirement period. So 3% was the right target all along.

So a $1m target for $40K a year spend turns into $1.33M. If the 3% SWR worked in 1929 and 1966 then for 50 years (or whatever) unless you retired into a cohort that experiences worse than 1929 and 1966 then $1.33M is safe. Hence the name.

It is certainly faster to get from $1M to $1.33M when returns are good…as long as your luck holds you can hit your new target before the next correction to retire safely.

If not and the market crashes you go backwards and have wait for the market to recover.

Or the market stagnates and you don’t get a lot of help to get to $1.33M and have to save $333K the hard way.

You completely missed my point.

The post I was responding to was making the reasonable argument that you need 3% WR rather than 4% because of low expected forward returns.
Then in the next breath said it was easy to get to 3% in a short few years because good market returns would do most of the work for you.

My point was not against needing 3%, but rather in the trivializing what it might take to get there.
If you are expecting poor returns and want to get to a lower WR because of that, you have to consider the same low-return environment when estimating how long it will take you to get there.
I think you're also mischaracterizing what I said. I didn't say anybody could just work a minimum wage job at McDonald's, listen to Dave Ramsay, pick an actively managed mutual fund with a proven track record of outperforming the market by 10% or more, wait for a few years, and retire a billionaire. I said that assuming someone can keep a presumably good-paying job that got them to FIRE levels of wealth at a young age for a few more years, there's a chance that between new contributions and something like a 5% real return (below the historical US average to account for high valuations, and again with the major caveat that the distribution of outcomes is wide regardless), they could reach that 3% level in a few years. They may also lose their job in the recession, the bear market could last an unusually long time, inflation may increase their cost of living faster than the portfolio can keep up, their country might get in a war, etc. Unfortunately, there is no strategy for common people to avoid the absolute worst outcomes other than becoming so wealthy that you can afford a nuclear bomb shelter or a ticket to a colony on Mars (not to mention becoming that rich carries its own risks). But if you reach a 4% SWR and want to reach 3% to be extra safe or support a very long retirement, you need to work some more, plain and simple. I don't know how long exactly that will take for each individual. 6 years was a really, really rough estimate based on what I thought was a reasonable growth projection, and then subtracting some time to account for new contributions could bring that down to 3-5 years or so. I don't think I'm saying anything crazy here, or that saying some people might get screwed and end up taking way longer is really all that helpful. At the end of the day, nobody really knows what they need or how to get there. Boglehead philosophy is based on time-tested strategies, but by time-tested, we really only mean a couple of centuries at most, not all of human or Earth history. For all we know, the Earth gets conquered by space aliens from another galaxy, and their first action as new rules of our planet is to outlaw publicly traded corporations and put all investors in jail, in which case none of these strategies are going to work. I think the projections I gave were in the ballpark of likely reality, but the usual disclaimer applies: "past performances does not guarantee future results".
marcopolo
Posts: 6469
Joined: Sat Dec 03, 2016 10:22 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by marcopolo »

dboeger1 wrote: Fri Aug 05, 2022 3:23 pm
marcopolo wrote: Fri Aug 05, 2022 2:02 pm
nigel_ht wrote: Fri Aug 05, 2022 1:44 pm
marcopolo wrote: Fri Aug 05, 2022 1:13 pm You see no irony in saying people should work longer to get to 3% because poor returns are likely when markets are near their highs, and then saying it should only take a few years to get there because of the good market returns?!?
Nope. I’m not the op but it makes sense.

Most FIRE folks need 33x vs 25x because they have a longer retirement period. So 3% was the right target all along.

So a $1m target for $40K a year spend turns into $1.33M. If the 3% SWR worked in 1929 and 1966 then for 50 years (or whatever) unless you retired into a cohort that experiences worse than 1929 and 1966 then $1.33M is safe. Hence the name.

It is certainly faster to get from $1M to $1.33M when returns are good…as long as your luck holds you can hit your new target before the next correction to retire safely.

If not and the market crashes you go backwards and have wait for the market to recover.

Or the market stagnates and you don’t get a lot of help to get to $1.33M and have to save $333K the hard way.

You completely missed my point.

The post I was responding to was making the reasonable argument that you need 3% WR rather than 4% because of low expected forward returns.
Then in the next breath said it was easy to get to 3% in a short few years because good market returns would do most of the work for you.

My point was not against needing 3%, but rather in the trivializing what it might take to get there.
If you are expecting poor returns and want to get to a lower WR because of that, you have to consider the same low-return environment when estimating how long it will take you to get there.
I think you're also mischaracterizing what I said. I didn't say anybody could just work a minimum wage job at McDonald's, listen to Dave Ramsay, pick an actively managed mutual fund with a proven track record of outperforming the market by 10% or more, wait for a few years, and retire a billionaire. I said that assuming someone can keep a presumably good-paying job that got them to FIRE levels of wealth at a young age for a few more years, there's a chance that between new contributions and something like a 5% real return (below the historical US average to account for high valuations, and again with the major caveat that the distribution of outcomes is wide regardless), they could reach that 3% level in a few years. They may also lose their job in the recession, the bear market could last an unusually long time, inflation may increase their cost of living faster than the portfolio can keep up, their country might get in a war, etc. Unfortunately, there is no strategy for common people to avoid the absolute worst outcomes other than becoming so wealthy that you can afford a nuclear bomb shelter or a ticket to a colony on Mars (not to mention becoming that rich carries its own risks). But if you reach a 4% SWR and want to reach 3% to be extra safe or support a very long retirement, you need to work some more, plain and simple. I don't know how long exactly that will take for each individual. 6 years was a really, really rough estimate based on what I thought was a reasonable growth projection, and then subtracting some time to account for new contributions could bring that down to 3-5 years or so. I don't think I'm saying anything crazy here, or that saying some people might get screwed and end up taking way longer is really all that helpful. At the end of the day, nobody really knows what they need or how to get there. Boglehead philosophy is based on time-tested strategies, but by time-tested, we really only mean a couple of centuries at most, not all of human or Earth history. For all we know, the Earth gets conquered by space aliens from another galaxy, and their first action as new rules of our planet is to outlaw publicly traded corporations and put all investors in jail, in which case none of these strategies are going to work. I think the projections I gave were in the ballpark of likely reality, but the usual disclaimer applies: "past performances does not guarantee future results".
Perhaps in that case you don't fully appreciate how conservative a 4% WR is, and what it takes for it to fail.

If you make reasonable growth projections, then 4% WR works fine.
If you think growth will be significantly less (really significantly less to put 4% WR at risk), then you should really use that same really poor growth projection in estimating how long one would need to work to get to a 3% WR.

If we get anywhere near the 5% real returns you use in saying it only takes a few years to go from 4% to 3% WR, there would be no reason to do that.
At 5% real return, a 4% WR would leave the retiree with a huge pile of money left over.
You need something around 1% real to put 4% WR at risk. Now, how long does it take to go from 4% to 3% WR?

The real risk to 4% WR has typically been due to SORR showing up.
Again, making the assumption of 5% real for the first 6 years essentially eliminates SORR as a concern.
If you consider cases where SORR does indeed show up (so you need lower WR), now how long do you need to go from 4% to 3%?
Once in a while you get shown the light, in the strangest of places if you look at it right.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by willthrill81 »

marcopolo wrote: Fri Aug 05, 2022 12:40 pm
dboeger1 wrote: Thu Aug 04, 2022 4:26 pm
Tamalak wrote: Thu Aug 04, 2022 9:06 am What I don't like about these "odds your money will last" sites is they seem to assume that you're equally likely to retire every year.

But people are much more likely to hit their number when the market was high. Nobody hit it at the bottom of 2008. So I feel like it might be overly optimistic.
This is an excellent point. I've been reading ERN's SWR series recently, and one of the points he frequently mentions is that because stock returns are correlated with CAPE ratio, you may have to adjust SWR lower based on current market conditions. It's easy to get tricked into thinking you have enough money when your paper wealth is inflated by high prices, but that doesn't mean you'll actually get average real returns starting from that point (to be fair, the distribution of outcomes was very wide in all CAPE bands, it just shifted better or worse). So for most people, something like the 4% rule should probably be renamed the "most likely 3.5% but maybe 4% or higher depending on valuations when you retire" rule. Doesn't roll off the tongue quite as well though, does it?

I think you touched on an interesting point too, which is that people like to come up with "their number", but having a CAPE-dependent SWR suggests that there is no single number. The number needs to be adjusted for reasonable equity return assumptions based on current valuations.

I'm increasingly becoming convinced that 3% SWR is a much more reasonable target than 4%, particularly for people on the earlier end of the retirement spectrum. It's sad, but I suppose it's better to end up with more money than to run out of it in retirement. I can see why ERN has been called The Grinch of Early Retirement.

The good news is that it shouldn't take very long to reach 3% SWR from 4%; the required balance is only 1/3 more. With absolutely no new contributions and a 5% real rate of return, it takes just shy of 6 years. Obviously, the exact time depends greatly on market returns and other factors, but with new contributions added into the mix, I don't think it's crazy to assume that difference could be made up in 3-4 years of additional work. Now, is that always a valid option for everybody? No, some people are forced out of work, work in toxic environments, choose to become stay-at-home parents, etc. I don't want to make it sound like it's nothing. But for people of means on the edge of retirement, I think it's comforting to know that just 3-4 more years of work can go from a 4% SWR with a lot of assumptions built in to a 3% SWR which has historically supported very long retirements through pretty rough market conditions. Just something to consider.
You see no irony in saying people should work longer to get to 3% because poor returns are likely when markets are near their highs, and then saying it should only take a few years to get there because of the good market returns?!?
Precisely. Investors who were at 25x in the year 2000 and were saving 20% of their income into a 60/40 AA wouldn't have reached 33x until 2012. Something like that is what the '3% is the new 4%' folks need to be preparing for if they think that forward returns are so poor that 4% is too high of a starting withdrawal rate.
User avatar
unclescrooge
Posts: 5984
Joined: Thu Jun 07, 2012 7:00 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by unclescrooge »

billaster wrote: Thu Aug 04, 2022 1:06 pm It's sort of like insisting that airplanes be 100% safe when you are more likely to die driving to the airport.
The only reason you're more likely to die driving to the airport is because airplanes are so safe.
dboeger1
Posts: 1259
Joined: Fri Jan 13, 2017 7:32 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by dboeger1 »

marcopolo wrote: Fri Aug 05, 2022 3:53 pm
dboeger1 wrote: Fri Aug 05, 2022 3:23 pm
marcopolo wrote: Fri Aug 05, 2022 2:02 pm
nigel_ht wrote: Fri Aug 05, 2022 1:44 pm
marcopolo wrote: Fri Aug 05, 2022 1:13 pm You see no irony in saying people should work longer to get to 3% because poor returns are likely when markets are near their highs, and then saying it should only take a few years to get there because of the good market returns?!?
Nope. I’m not the op but it makes sense.

Most FIRE folks need 33x vs 25x because they have a longer retirement period. So 3% was the right target all along.

So a $1m target for $40K a year spend turns into $1.33M. If the 3% SWR worked in 1929 and 1966 then for 50 years (or whatever) unless you retired into a cohort that experiences worse than 1929 and 1966 then $1.33M is safe. Hence the name.

It is certainly faster to get from $1M to $1.33M when returns are good…as long as your luck holds you can hit your new target before the next correction to retire safely.

If not and the market crashes you go backwards and have wait for the market to recover.

Or the market stagnates and you don’t get a lot of help to get to $1.33M and have to save $333K the hard way.

You completely missed my point.

The post I was responding to was making the reasonable argument that you need 3% WR rather than 4% because of low expected forward returns.
Then in the next breath said it was easy to get to 3% in a short few years because good market returns would do most of the work for you.

My point was not against needing 3%, but rather in the trivializing what it might take to get there.
If you are expecting poor returns and want to get to a lower WR because of that, you have to consider the same low-return environment when estimating how long it will take you to get there.
I think you're also mischaracterizing what I said. I didn't say anybody could just work a minimum wage job at McDonald's, listen to Dave Ramsay, pick an actively managed mutual fund with a proven track record of outperforming the market by 10% or more, wait for a few years, and retire a billionaire. I said that assuming someone can keep a presumably good-paying job that got them to FIRE levels of wealth at a young age for a few more years, there's a chance that between new contributions and something like a 5% real return (below the historical US average to account for high valuations, and again with the major caveat that the distribution of outcomes is wide regardless), they could reach that 3% level in a few years. They may also lose their job in the recession, the bear market could last an unusually long time, inflation may increase their cost of living faster than the portfolio can keep up, their country might get in a war, etc. Unfortunately, there is no strategy for common people to avoid the absolute worst outcomes other than becoming so wealthy that you can afford a nuclear bomb shelter or a ticket to a colony on Mars (not to mention becoming that rich carries its own risks). But if you reach a 4% SWR and want to reach 3% to be extra safe or support a very long retirement, you need to work some more, plain and simple. I don't know how long exactly that will take for each individual. 6 years was a really, really rough estimate based on what I thought was a reasonable growth projection, and then subtracting some time to account for new contributions could bring that down to 3-5 years or so. I don't think I'm saying anything crazy here, or that saying some people might get screwed and end up taking way longer is really all that helpful. At the end of the day, nobody really knows what they need or how to get there. Boglehead philosophy is based on time-tested strategies, but by time-tested, we really only mean a couple of centuries at most, not all of human or Earth history. For all we know, the Earth gets conquered by space aliens from another galaxy, and their first action as new rules of our planet is to outlaw publicly traded corporations and put all investors in jail, in which case none of these strategies are going to work. I think the projections I gave were in the ballpark of likely reality, but the usual disclaimer applies: "past performances does not guarantee future results".
Perhaps in that case you don't fully appreciate how conservative a 4% WR is, and what it takes for it to fail.

If you make reasonable growth projections, then 4% WR works fine.
If you think growth will be significantly less (really significantly less to put 4% WR at risk), then you should really use that same really poor growth projection in estimating how long one would need to work to get to a 3% WR.

If we get anywhere near the 5% real returns you use in saying it only takes a few years to go from 4% to 3% WR, there would be no reason to do that.
At 5% real return, a 4% WR would leave the retiree with a huge pile of money left over.
You need something around 1% real to put 4% WR at risk. Now, how long does it take to go from 4% to 3% WR?

The real risk to 4% WR has typically been due to SORR showing up.
Again, making the assumption of 5% real for the first 6 years essentially eliminates SORR as a concern.
If you consider cases where SORR does indeed show up (so you need lower WR), now how long do you need to go from 4% to 3%?
Okay, but understand that the whole reason for the 3% target was so that you can make it through bad luck SORR immediately after retiring with a 3% WR. Yes, if SORR shows up while on the way to 3%, it's going to make it harder to hit 3% because the portfolio is dropping... but the SWR is also rising because the valuations are no longer so high and the risk of poor future returns is reduced. It's not conceptually difficult to understand, there is some inherent value in stocks to which prices are very loosely anchored. You generally don't expect the market to keep dropping forever despite companies continuing to generate earnings. So yeah, you may not reach 3% if get a bear market, but you don't necessarily have to either. 3% WR in 2009 would've been ridiculously conservative given what the market did over the next decade. I get the point you're trying to make, but it just doesn't change the recommendation all that much. Objectively, continuing to work and save will almost always improve one's retirement prospects by both increasing assets and reducing retirement length (I believe one could technically come up with a contrived example where retiring and spending early in a healthier market before an extreme prolonged downturn would be better than retiring in that downturn, but that's pretty extreme). You're making it sound like my suggestion to target 3% is going to result in people working for decades and never getting there. I just don't see it that way, unless someone is religious about the 3% target in the face of economic and market conditions that suggest another target might be more appropriate or reasonable.
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 47214
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Watch out for hidden risk tolerance assumptions in SWR cla

Post by nisiprius »

PicassoSparks wrote: Fri Aug 05, 2022 10:20 am
nigel_ht wrote: Fri Aug 05, 2022 9:32 am Um...that it is a ceiling implies that your actual WR is likely lower most of the time...which isn't inflexible.
No. The SWR that's used for the calculations of the Trinity study and the charts we saw upthread is extremely mechanical...
Not really. The SWR used for the calculation may be mechanical, but the words of the authors and their advice to retirees are not. From the article:
The word planning is emphasized because of the great uncertainties in the stock and bond markets. Mid-course corrections likely will be required, with the actual dollar amounts withdrawn adjusted downward or upward relative to the plan. The investor needs to keep in mind that selection of a withdrawal rate is not a matter of contract but rather a matter of planning.
I had a question and I got the senior author's email address from somewhere, possibly the university's online directory, and wrote to him:
What the "4% SWR" means is not that you can treat a portfolio as if it were a guaranteed annuity.

I think all the [Trinity] authors meant is that if it is late 2008 and your stocks halve in value, you don't need to halve your spending instantly. It's OK to cross your fingers and continue spending according to the 4%-then-COLAed plan, even though it means dipping into capital, and it's OK to go on doing that for a while.
Professor Cooley replied:
You have hit the nail on the head! I've tried to explain that thought to journalists but they don't seem to get it. You've got it. Stay flexible my friend!, which is the advice we should give to retirees.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
User avatar
PicassoSparks
Posts: 334
Joined: Tue Apr 28, 2020 5:41 am

Re: Watch out for hidden risk tolerance assumptions in SWR cla

Post by PicassoSparks »

nisiprius wrote: Fri Aug 05, 2022 6:11 pm
PicassoSparks wrote: Fri Aug 05, 2022 10:20 am No. The SWR that's used for the calculations of the Trinity study and the charts we saw upthread is extremely mechanical...
Not really. The SWR used for the calculation may be mechanical, but the words of the authors and their advice to retirees are not. From the article:
The word planning is emphasized because of the great uncertainties in the stock and bond markets. Mid-course corrections likely will be required, with the actual dollar amounts withdrawn adjusted downward or upward relative to the plan.
I think we’re agreeing with each other. What I’m trying to convey is an understanding that a mechanical WR rate requires you to have a much lower starting rate than if you are willing to adjust your WR in the face of disappointing returns, enabling you to compound that reduction. While at the same time, you don’t need to have your withdrawal rate jump up and down as your investments jump around with volatility.

The main point I take from your starting post is that managing WR is substantially more important than AA.
Zeno
Posts: 769
Joined: Wed Sep 12, 2018 10:44 am

Re: Watch out for hidden risk tolerance assumptions in SWR cla

Post by Zeno »

Great insights and information as always, nisiprius. Thank you.
Last edited by Zeno on Sun Aug 07, 2022 8:12 pm, edited 1 time in total.
Northern Flicker
Posts: 11031
Joined: Fri Apr 10, 2015 12:29 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Northern Flicker »

saver7007 wrote: Fri Aug 05, 2022 2:48 pm
nisiprius wrote: Fri Aug 05, 2022 6:35 am In other words, the required planning time frame for an individual is not-quite-double average life expectancy.
That would make a pretty awesome rule of thumb for (early) retirement planning, I just went and looked this up for myself.
How does average life expectancy differ from life expectancy?

If the planning horizon is almost double life expectancy, that would make SPIAs very cost effective, as they are priced to fund only up to life expectancy.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
saver7007
Posts: 40
Joined: Thu Jan 06, 2022 2:30 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by saver7007 »

Northern Flicker wrote: Sat Aug 06, 2022 1:59 am
saver7007 wrote: Fri Aug 05, 2022 2:48 pm
nisiprius wrote: Fri Aug 05, 2022 6:35 am In other words, the required planning time frame for an individual is not-quite-double average life expectancy.
That would make a pretty awesome rule of thumb for (early) retirement planning, I just went and looked this up for myself.
How does average life expectancy differ from life expectancy?

If the planning horizon is almost double life expectancy, that would make SPIAs very cost effective, as they are priced to fund only up to life expectancy.
I'm not an actuary and not the right guy to define any terms related to life expectancy, but the tables here give "the average remaining number of years expected prior to death for a person at [their] exact age".

An SPIA does seem like a very cost effective way to cover longevity risk in principle (as long as the fees and expenses are low). The cost of an SPIA should be much lower than the necessary size of an individual portfolio that covers longevity risk for any given stream of payments.
JackoC
Posts: 3623
Joined: Sun Aug 12, 2018 11:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by JackoC »

nigel_ht wrote: Thu Aug 04, 2022 12:46 pm
JackoC wrote: Thu Aug 04, 2022 12:16 pm
nigel_ht wrote: Thu Aug 04, 2022 11:37 am
Well, most modern planning tools allow for a 100% historical success rate which is what I use…and 4% is better than 3.9something from a simplicity standpoint.
I applaud you for noting that what the historically base simulations give you is the *historical* probability.
1. Yes…and that’s all SWR promises.
But I assume OP is talking about the forward looking probability.
2. I hope not, and I don’t think that’s correct given the OP.
As in example I just gave, if I rig up FireCalc with a generic 98% historical success plan. I believe the actual probability of success of that plan is more like 2/3's. I'm not certain it's exactly 2/3's but no way would I plan thinking it was 98%, looking in the direction that matters, forward.
3. If I thought that using a SWR number with a 98% historical success probability had only a 66% real world chance of success I’d probably buy a lot more gold.

4. 33% risk of failure would require me to hold so much management reserve that the numbers simply wouldn’t work out.
1. Then it's not worth much. There has to be some assumption that the historical probability is quantitatively related to the future probability to make the historical probability any serious part of a future plan. But there's no good reason for that assumption IMO.
2. It's obvious from the context of the whole discussion, whole theme of the forum, we're talking about planning for the future, not a simple examination of past financial history. There must be some assumption the past probability (independent of what the valuations and yields were at starting points in the past sample) is quantitatively predictive. But I don't think so.
3. I don't see why. Simply, if the mean of the future return distribution is 2% pa lower than the mean of the past return distribution and the variance and higher moments are the same, that's roughly what it could come out. However as I pointed out in the same post, cutting the WR to only 3.6% from 4% under the same return assumption raises the 30yr success probability from 66% to 83% and shooting for forward looking (translation, 'relevant') probability of much more than 80% is probably pointless.
4. This isn't a good reason to reject realistic forward looking return estimates in planning for the future IMO, 'but that would spoil my preconceived plan'. But it does seem to factor in as a reason in many cases.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by willthrill81 »

Northern Flicker wrote: Sat Aug 06, 2022 1:59 am If the planning horizon is almost double life expectancy, that would make SPIAs very cost effective, as they are priced to fund only up to life expectancy.
The problem with SPIAs these days is that none of them are CPI linked anymore. This means that if the longevity risk one is buying them to insure against shows up, inflation is likely to significantly reduce their buying power. 30 years of 2% inflation leaves you with only 55% of the buying power you started with, and 3% inflation would knock that down to only 40%.

This problem can be partly resolved by adding something like a 2-3% fixed COLA, but its usefulness depends on future inflation being relatively close to your COLA. A few years of inflation like we're seeing now could take a big and permanent bite out of a SPIA payout's buying power.
Broken Man 1999
Posts: 7483
Joined: Wed Apr 08, 2015 11:31 am
Location: West coast of Florida, near Champa Bay !

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Broken Man 1999 »

nisiprius wrote: Fri Aug 05, 2022 9:04 am
vineviz wrote: Fri Aug 05, 2022 8:24 am...Upshot: a health 65 year old couple should almost surely be planning for at least a 30 year retirement. And as they age, they will continually (but gradually) push out the plan such that if both are alive and health at age 90 they are planning for a solid chance that at least one of them will live 8-10 more years...
Yes. It reminds me seriously of the time I was perusing a copy of the World Almanac and Book of Facts* at perhaps age ten. It had a table of life expectancies for each age. I asked my mom to explain it. She said your life expectancy was the number of years you could reasonably expect to live at that age. So I interpreted recursively... I figured at birth you could "expect" to live 70 years (or whatever the number was then), but at age 70 you could "expect" to live another 15 years, so you but at age 85 you could still "expect" to live another 6 years... and at age 91, another 3... so I thought it was saying that your life expectancy at birth was really 94 years.



*Ah, the World Almanac. Rendered obsolete by the Internet, but what a wonderful thing it once was. Still in print, I guess. Printed on newsprint, about the size of a trade paperback but about 1.5 inches thick, and the most clever selection of "facts..." "World" didn't mean "global," it meant the name of a newspaper... that became the "World-Telegram" and then the "World-Telegram and Sun" and then the "World Journal Tribune" and then nothing.
World Almanac was present in our home. I also liked the Statistical Abstract of the United States.

Broken Man 1999
“If I cannot drink Bourbon and smoke cigars in Heaven then I shall not go." - Mark Twain
JackoC
Posts: 3623
Joined: Sun Aug 12, 2018 11:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by JackoC »

dboeger1 wrote: Fri Aug 05, 2022 1:32 pm Ultimately, the universe doesn't care how long it takes to reach 3% WR from 4%. It might take one person 2 years, and another 20 years. Either way, if 4% isn't low enough, then it isn't low enough. It's up to each person to decide the tradeoffs they want to make. If you were to ask me personally whether I'd work 20 years to drop from 4% to 3% WR, I would probably say no. I think I could most likely make 4% work, and I also have plans to make some income in early retirement through passion projects and small businesses which I have not built into my financial assumptions. But would I work 2 years to reach 3%? Most likely yes. I don't want to work another 2 years, but the difference in peace of mind would be huge. Again, the only point I was really trying to make is that 4% is likely not as safe as many people realize, and it might be worth working a little longer to reach that 3%.
Right, the market doesn't owe us anything. But the discussion sometimes seems like the 'negotiation' stage of grief :happy 'What can you give me to show it's practical to reach savings that would actually give me a high *forward looking* probability of success per your return estimate, or else your return estimate must be wrong, and you're a Cassandra, a gold bug or something.' But return estimates per fundamentals like div yield and plausible EPS growth rate (given GDP growth trend and long term historical lag of EPS growth well behind GDP), CAEY (1/CAPE), the long term riskless rate plus Equity Risk Premium (and looking at things like risky bond credit spreads ERP is probably also compressed from historical)...those things really don't care about you and me. And if future variance and higher moments (fatness of tails etc) are not lower than past, it means in simple statistics future bad cases will be more common, and yes failure rate at what achieved 2% failure in past will fail a large multiple more often (that's how the shape of fatish/skewed normal-ish distribution tails work). Or maybe you can argue forward looking variance is lower than past, or the left tail is somehow bounded by what happened in 3-5 independent 30 yr samples in the past in one country. . But it's not necessarily easy to separate that from the impact on personal plans already made with more optimistic assumptions.

I'd also note again though if one crudely corrects FireCalc to a more reasonable future mean return by putting in 2% pa dummy expense (others could argue 2% isn't enough, just for illustration) a generic 98% successful 30 yr 4% becomes 66%. But at 3.6% isn't already back to 83%, 3% it's back to 98%. For the traditional-type retirement length having the priniciple to burn through significantly insulates against bad returns, besides it being debatable if 98% is a reasonable minimum or something more like 80 is more realistic in the general context of end of life. Where using past returns leads to really seriously deluded plans IMO is in very early retirement or statements about the 'perpetual SWR'. And everybody knows (though I'll repeat it anyway) SWR (initial % then same real $ amount thereafter) is a simple comparative index of savings adequacy not a tactic for actual withdrawals. But '4% SWR, per the stated rule has 98% chance of success' is an analytical not personal statement, and one which is seriously distorted in an optimistic direction on a forward looking basis IMO. 'I believe the underlying return assumption in that analysis is incorrect', the statement is not 'you will stick to a fixed inflation adjusted withdrawal for 100% of your wealth and end up surviving on public/family aid/charity X% of the time'.
Last edited by JackoC on Sat Aug 06, 2022 10:08 am, edited 1 time in total.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

JackoC wrote: Sat Aug 06, 2022 8:46 am
nigel_ht wrote: Thu Aug 04, 2022 12:46 pm
JackoC wrote: Thu Aug 04, 2022 12:16 pm
nigel_ht wrote: Thu Aug 04, 2022 11:37 am
Well, most modern planning tools allow for a 100% historical success rate which is what I use…and 4% is better than 3.9something from a simplicity standpoint.
I applaud you for noting that what the historically base simulations give you is the *historical* probability.
1. Yes…and that’s all SWR promises.
But I assume OP is talking about the forward looking probability.
2. I hope not, and I don’t think that’s correct given the OP.
As in example I just gave, if I rig up FireCalc with a generic 98% historical success plan. I believe the actual probability of success of that plan is more like 2/3's. I'm not certain it's exactly 2/3's but no way would I plan thinking it was 98%, looking in the direction that matters, forward.
3. If I thought that using a SWR number with a 98% historical success probability had only a 66% real world chance of success I’d probably buy a lot more gold.

4. 33% risk of failure would require me to hold so much management reserve that the numbers simply wouldn’t work out.
1. Then it's not worth much. There has to be some assumption that the historical probability is quantitatively related to the future probability to make the historical probability any serious part of a future plan. But there's no good reason for that assumption IMO.

2. It's obvious from the context of the whole discussion, whole theme of the forum, we're talking about planning for the future, not a simple examination of past financial history. There must be some assumption the past probability (independent of what the valuations and yields were at starting points in the past sample) is quantitatively predictive. But I don't think so.
Historical performance tells you that for the worst case that X% WR was sustainable. 100% success rate isn't predictive of future performance but mathematical.

It is up to you to determine the probability that you will experience a worse than historical worst case scenario.

The value of SWR is it gives you a (conservative) starting point for your planning. This plan survived 1929 and 1966. Good.

Now what are the most likely risks of failure given this starting point and how much additional mitigation should I apply?

It's like flood planning. The worst case historical scenario was a 14 foot storm surge in 1966...so the sea wall should be no less than 14'. That's the SWR starting point...14 feet had 100% historical success rate.

Now taking into account subsidence, sea level rise, localized erosion, observed increase in storm strength and increase in frequency of "100 year storms" how high should the sea wall be?

That depends on your analysis, risk tolerance and available budget.

But historical data is by no means "not worth much".
3. I don't see why. Simply, if the mean of the future return distribution is 2% pa lower than the mean of the past return distribution and the variance and higher moments are the same, that's roughly what it could come out. However as I pointed out in the same post, cutting the WR to only 3.6% from 4% under the same return assumption raises the 30yr success probability from 66% to 83% and shooting for forward looking (translation, 'relevant') probability of much more than 80% is probably pointless.

4. This isn't a good reason to reject realistic forward looking return estimates in planning for the future IMO, 'but that would spoil my preconceived plan'. But it does seem to factor in as a reason in many cases.
Nope. This is saying that the additional sea wall height to get to 100% success rate is cost prohibitive.

That's what you're asserting: a withdrawal rate that survived 1929/1966 has only a 66% future success rate means that 33% of the future cohorts will fail.

This assertion is equivalent to asserting a 14 foot sea wall, which historically stopped all storm surges, will only likely defend the coastline 66% of the time from future storms. That of the next 100 years, 33 will experience "100 year storms" that produce greater 14' storm surges. That "100 year storms" will become commonplace.

If this assessment is "realistic" then the "realistic" response is "Gee, maybe I should stop living on the coastline".
User avatar
vineviz
Posts: 12280
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by vineviz »

nigel_ht wrote: Sat Aug 06, 2022 10:06 am The value of SWR is it gives you a (conservative) starting point for your planning. This plan survived 1929 and 1966. Good.
Maybe.

Without further analysis it's really difficult to know whether the historical SWR is a "conservative" or "aggressive" starting point.

In other words, it's only "conservative" if we assume, implicitly or explicitly, that the distribution of real returns hasn't shifted in any important ways.

The challenge: we can clearly observe that such shifts HAVE occurred.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
JackoC
Posts: 3623
Joined: Sun Aug 12, 2018 11:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by JackoC »

nigel_ht wrote: Sat Aug 06, 2022 10:06 am
JackoC wrote: Sat Aug 06, 2022 8:46 am
nigel_ht wrote: Thu Aug 04, 2022 12:46 pm
JackoC wrote: Thu Aug 04, 2022 12:16 pm
nigel_ht wrote: Thu Aug 04, 2022 11:37 am Well, most modern planning tools allow for a 100% historical success rate which is what I use…and 4% is better than 3.9something from a simplicity standpoint.
I applaud you for noting that what the historically base simulations give you is the *historical* probability.
1. Yes…and that’s all SWR promises.
But I assume OP is talking about the forward looking probability.
2. I hope not, and I don’t think that’s correct given the OP.
As in example I just gave, if I rig up FireCalc with a generic 98% historical success plan. I believe the actual probability of success of that plan is more like 2/3's. I'm not certain it's exactly 2/3's but no way would I plan thinking it was 98%, looking in the direction that matters, forward.
3. If I thought that using a SWR number with a 98% historical success probability had only a 66% real world chance of success I’d probably buy a lot more gold.

4. 33% risk of failure would require me to hold so much management reserve that the numbers simply wouldn’t work out.
1. Then it's not worth much. There has to be some assumption that the historical probability is quantitatively related to the future probability to make the historical probability any serious part of a future plan. But there's no good reason for that assumption IMO.

2. It's obvious from the context of the whole discussion, whole theme of the forum, we're talking about planning for the future, not a simple examination of past financial history. There must be some assumption the past probability (independent of what the valuations and yields were at starting points in the past sample) is quantitatively predictive. But I don't think so.
1. Historical performance tells you that for the worst case that X% WR was sustainable. 100% success rate isn't predictive of future performance but mathematical.

It is up to you to determine the probability that you will experience a worse than historical worst case scenario.

2. The value of SWR is it gives you a (conservative) starting point for your planning. This plan survived 1929 and 1966. Good.

Now what are the most likely risks of failure given this starting point and how much additional mitigation should I apply?

3. It's like flood planning. The worst case historical scenario was a 14 foot storm surge in 1966...so the sea wall should be no less than 14'. That's the SWR starting point...14 feet had 100% historical success rate.

Now taking into account subsidence, sea level rise, localized erosion, observed increase in storm strength and increase in frequency of "100 year storms" how high should the sea wall be?

That depends on your analysis, risk tolerance and available budget.

4. But historical data is by no means "not worth much".
3. I don't see why. Simply, if the mean of the future return distribution is 2% pa lower than the mean of the past return distribution and the variance and higher moments are the same, that's roughly what it could come out. However as I pointed out in the same post, cutting the WR to only 3.6% from 4% under the same return assumption raises the 30yr success probability from 66% to 83% and shooting for forward looking (translation, 'relevant') probability of much more than 80% is probably pointless.

4. This isn't a good reason to reject realistic forward looking return estimates in planning for the future IMO, 'but that would spoil my preconceived plan'. But it does seem to factor in as a reason in many cases.
5. Nope. This is saying that the additional sea wall height to get to 100% success rate is cost prohibitive.

6. If this assessment is "realistic" then the "realistic" response is "Gee, maybe I should stop living on the coastline".
1. Tells you what *was* the worst case based on starting yields and valuations then prevailing in a quite small sample of *non-overlapping* (not autocorrelated) 30 yr periods. It is up to us to examine this assumption I agree, but I believe we can find a more accurate estimate if we incorporate *now's* valuations/yields as the starting point. The answer will then change quite a bit in the tails (though as noted twice already the SWR answer for a period like 30 yrs doesn't change dramatically because it has the principle to burn through).

2. Not conservative given now's valuations and yields, as opposed to unconditional selection from a past distribution characterized by lower valuation/higher yield.

3. It is, and remarkable you'd use that as supposedly supporting your point. If (in a hypothetical world, I am *not* debating climate change on real earth) sea level was rising significantly, even with big storms only as likely (but no more likely) than in the last 100yrs, the '100 yr' floodplain based on history could not be taken literally for future planning purposes. That's all I'm doing here, adjusting the probability for the higher sea level (lower expected return). If we assumed big storms were more likely in the future it would be worse. If the middle of the forward looking return distribution is as much lower than the past geometric mean as the expected return now is, and future variance/higher moments of return are the same, the machine will spit out 66% rather than 98% (a properly constructed simulation using now's mean and the past variance/higher moments, rather than crude correction of 2% expenses isn't going to show exactly 66% but it will be much lower than 98%).

4. It's not that no historical data is worth anything. Thers are specific statistical problems with how FireCalc et work. They assume return is independent of valuation by equally weighting past cases regardless of starting valulation lower (vast majority of cases) or higher than now. They use overlapping periods to give an impression of robust data, but it's actually pretty little data about long period returns (the overlapping paths are various reruns of the same few macro facts, Great Depression didn't bring down the whole system, US on winning side of the WW's, inflation of the 60's was conquered when it was, etc.). And it's ignoring current market data, the yield on bonds now for example if not a direct input at all. So it's irrelevant to my return on 60/40 for 30 yrs what yield I buy the 30 yr TIPS at now? That doesn't make sense. Using historical data in some shape or form is inevitable for some things, but here it's being used in an incorrect way IMO.

5. The logical answer to that is not to reject the evidence that sea level will rise (on the *hypothetical planet*) or to assume storms will be milder in the future than past (so worst case storm surge from a higher mean sea level will only be as high as past worst case).

6. But no such choice exists in the retirement planning case. We must realistically assess expected return, variance and higher moments on a *forward looking basis* as best we can and adjust savings rate, timing of retirement, consumption in retirement and certainty of sustaining
that consumption to make the numbers work. Again I gave simple example, FireCalc says 4% is 98% likely to work for 30 yrs. A more realistic return assumption from now with same variance/higher moments might get 3.6% 83% likely to work for 30 yrs. That's disappointing if one really wants to believe 4% is 98% likely to work, but I don't see how it's unworkable. It's not that much lower WR and still a high chance of success in the context of general uncertainties of end of life.
Last edited by JackoC on Sat Aug 06, 2022 10:55 am, edited 2 times in total.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

vineviz wrote: Sat Aug 06, 2022 10:12 am
nigel_ht wrote: Sat Aug 06, 2022 10:06 am The value of SWR is it gives you a (conservative) starting point for your planning. This plan survived 1929 and 1966. Good.
Maybe.

Without further analysis it's really difficult to know whether the historical SWR is a "conservative" or "aggressive" starting point.

In other words, it's only "conservative" if we assume, implicitly or explicitly, that the distribution of real returns hasn't shifted in any important ways.

The challenge: we can clearly observe that such shifts HAVE occurred.
I dunno...basing your withdrawal amount on the historical WORST case certainly isn't aggressive. It's not like SWR is based on median performance.

If the assertion is correct then every strategy less conservative (ie supporting a higher withdrawal rate) than SWR will likely also have significant issues going forward...
billaster
Posts: 391
Joined: Wed Apr 27, 2022 2:21 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by billaster »

nisiprius wrote: Fri Aug 05, 2022 6:35 am In other words, the required planning time frame for an individual is not-quite-double average life expectancy.
"Required" is your personal opinion on risk aversion. It is not an objective, measurable fact.
User avatar
vineviz
Posts: 12280
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by vineviz »

nigel_ht wrote: Sat Aug 06, 2022 10:46 am
I dunno...basing your withdrawal amount on the historical WORST case certainly isn't aggressive. It's not like SWR is based on median performance.
Again, there’s an important assumption behind this statement that may not be (and probably isn’t) true.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Blue456
Posts: 1744
Joined: Tue Jun 04, 2019 5:46 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Blue456 »

billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.

So even with Peter Lynch's 7% withdrawal, which some call "unrealistic", your odds of being a broke 95-year-old are only somewhere around 5% to 10%.

It's sort of like insisting that airplanes be 100% safe when you are more likely to die driving to the airport.
Pascals Wager? It is much more painful be 85-95 eating cat food than dying at 85 with excess money.
billaster
Posts: 391
Joined: Wed Apr 27, 2022 2:21 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by billaster »

Blue456 wrote: Sat Aug 06, 2022 11:37 am
billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.

So even with Peter Lynch's 7% withdrawal, which some call "unrealistic", your odds of being a broke 95-year-old are only somewhere around 5% to 10%.

It's sort of like insisting that airplanes be 100% safe when you are more likely to die driving to the airport.
Pascals Wager? It is much more painful be 85-95 eating cat food than dying at 85 with excess money.
People misstate Pascal's wager all the time. He based it on the assumption that the possible gain is infinite while the cost is finite. That is not the case here.
Cash is King
Posts: 358
Joined: Sun Dec 10, 2017 9:04 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Cash is King »

Here's recent article from MMM on why you'll probably never run out of money:

https://www.mrmoneymustache.com/2022/07 ... -of-money/


I think the article makes some great points.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

vineviz wrote: Sat Aug 06, 2022 11:34 am
nigel_ht wrote: Sat Aug 06, 2022 10:46 am
I dunno...basing your withdrawal amount on the historical WORST case certainly isn't aggressive. It's not like SWR is based on median performance.
Again, there’s an important assumption behind this statement that may not be (and probably isn’t) true.
The assumption is that worse than worst case scenarios don't happen every decade...if it did I think index investing would be dead.

Given that we're talking about valuations then only 1929 applies...
User avatar
vineviz
Posts: 12280
Joined: Tue May 15, 2018 1:55 pm
Location: Baltimore, MD

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by vineviz »

nigel_ht wrote: Sat Aug 06, 2022 3:21 pm
The assumption is that worse than worst case scenarios don't happen every decade...if it did I think index investing would be dead.
No, the assumption is that the future will be the same as the past.

History's "worst case scenario" could be the future's "better than average" scenario. Don't assume it can't be.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch
Zeno
Posts: 769
Joined: Wed Sep 12, 2018 10:44 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Zeno »

vineviz wrote: Sat Aug 06, 2022 3:26 pm
nigel_ht wrote: Sat Aug 06, 2022 3:21 pm
The assumption is that worse than worst case scenarios don't happen every decade...if it did I think index investing would be dead.
No, the assumption is that the future will be the same as the past.

History's "worst case scenario" could be the future's "better than average" scenario. Don't assume it can't be.
+1

See, e.g., basically anything by Nassim Taleb.
Last edited by Zeno on Sun Aug 07, 2022 8:08 pm, edited 1 time in total.
Northern Flicker
Posts: 11031
Joined: Fri Apr 10, 2015 12:29 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Northern Flicker »

marcopolo wrote: Fri Aug 05, 2022 1:37 pm How does one get from 4% to 3% by "working a little longer"?
I assume you are focusing on the word "little". You can get from 4% to 3% by working longer if your the size of your portfolio increases from additional savings and appreciation.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
petulant
Posts: 2206
Joined: Thu Sep 22, 2016 1:09 pm

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by petulant »

billaster wrote: Sat Aug 06, 2022 12:18 pm
Blue456 wrote: Sat Aug 06, 2022 11:37 am
billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.

So even with Peter Lynch's 7% withdrawal, which some call "unrealistic", your odds of being a broke 95-year-old are only somewhere around 5% to 10%.

It's sort of like insisting that airplanes be 100% safe when you are more likely to die driving to the airport.
Pascals Wager? It is much more painful be 85-95 eating cat food than dying at 85 with excess money.
People misstate Pascal's wager all the time. He based it on the assumption that the possible gain is infinite while the cost is finite. That is not the case here.
It's not brought up as a restatement of the exact weights. It's brought up as an analogy that the magnitude of an outcome matters as much as the probability of it occurring. The application here is that being alive, old, and broke has a much larger negative magnitude than some other scenarios, which may or may not be true.
Random Poster
Posts: 2758
Joined: Wed Feb 03, 2010 10:17 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Random Poster »

Cash is King wrote: Sat Aug 06, 2022 12:48 pm Here's recent article from MMM on why you'll probably never run out of money:

https://www.mrmoneymustache.com/2022/07 ... -of-money/


I think the article makes some great points.
I think that the article uses “ridiculously contrived examples” to make its point, so much so that the underlying examples are largely so unrealistic that his point is equally in doubt.

Ultimately, when one elects to retire (early or otherwise), no one is going to bail that person out if things go south or don’t quite go according to someone else’s calculations of how things should go.

It is extremely easy to write about how someone will likely never run out of money. It is also quite easy to run almost infinite calculations that show the same thing.

But it is something else entirely to pull the plug yourself and actually do it, and rely on one’s portfolio.

As for me, at age 45, I guess that I’ve been retired for a little over two years now. I don’t miss work, but I miss some of the people I worked with, and it can be stressful to rely on my portfolio for the remainder of my life, even at my sub 2.5% withdrawal rate. I know that I certainly have an aversion to spending money and in some ways it has only gotten worse over time and with declining portfolio balances (even though the dividend and interest income from the portfolio have remained steady or slightly increased).

Having the apparent financial freedom to not have to work can, I suppose, be nice, but it can also be terribly terrifying for several reasons, and so I see no harm in someone wanting to increase their underlying balance to lower their withdrawal rate if doing so helps them feel more safe.

Because, as stated, no one is going to bail them out when things don’t go the way a calculator or some blogger says that they should.
Most experiences are better imagined.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by willthrill81 »

Blue456 wrote: Sat Aug 06, 2022 11:37 am
billaster wrote: Thu Aug 04, 2022 1:06 pm It seems rather odd to insist on, say, a 95% success rate for money lasting 30 years when the odds of a 65-year-old even being alive in 30 years is only 10%.

So even with Peter Lynch's 7% withdrawal, which some call "unrealistic", your odds of being a broke 95-year-old are only somewhere around 5% to 10%.

It's sort of like insisting that airplanes be 100% safe when you are more likely to die driving to the airport.
Pascals Wager? It is much more painful be 85-95 eating cat food than dying at 85 with excess money.
Beans and rice are far cheaper than cat food, which has increased in price by 12.5% over the last year. :D
Northern Flicker
Posts: 11031
Joined: Fri Apr 10, 2015 12:29 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Northern Flicker »

vineviz wrote: Sat Aug 06, 2022 10:12 am
nigel_ht wrote: Sat Aug 06, 2022 10:06 am The value of SWR is it gives you a (conservative) starting point for your planning. This plan survived 1929 and 1966. Good.
Maybe.

Without further analysis it's really difficult to know whether the historical SWR is a "conservative" or "aggressive" starting point.

In other words, it's only "conservative" if we assume, implicitly or explicitly, that the distribution of real returns hasn't shifted in any important ways.

The challenge: we can clearly observe that such shifts HAVE occurred.
While I generally agree, I will also offer a more optimistic alternative outlook.

Economists sometimes talk about the balance of savings and investment, which establishes the demand for bonds, as establishing real rates. It was often written that when boomers retire and start re-allocating from stocks to bonds, stocks will asca result do poorly. This did not materialize perhaps because young savers were buying stock at a rate sufficient to absorb the stock being shed by boomers. But the young savers did not have bonds to sell back to the boomers. So maybe the actual effect of boomers retiring was historically low interest rates due to the high demand for bonds (directly, or indirectly through funding of pension liabilities).

We may be getting close enough to the end of the cycle of boomers retiring, that maybe it is not an accident that there is evidence that we are entering a new secular phase for interest rates.

This could mean higher real portfolio returns in the future, and more in line with historical norms.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

vineviz wrote: Sat Aug 06, 2022 3:26 pm
nigel_ht wrote: Sat Aug 06, 2022 3:21 pm
The assumption is that worse than worst case scenarios don't happen every decade...if it did I think index investing would be dead.
No, the assumption is that the future will be the same as the past.
Where does it say that in any SWR studies?
History's "worst case scenario" could be the future's "better than average" scenario. Don't assume it can't be.
And if true there will be a lot less interest in the stock market...the 20 year period of 1929 to 1949 returned 2.11% real.
Northern Flicker
Posts: 11031
Joined: Fri Apr 10, 2015 12:29 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Northern Flicker »

willthrill81 wrote: Sat Aug 06, 2022 9:11 am
Northern Flicker wrote: Sat Aug 06, 2022 1:59 am If the planning horizon is almost double life expectancy, that would make SPIAs very cost effective, as they are priced to fund only up to life expectancy.
The problem with SPIAs these days is that none of them are CPI linked anymore. This means that if the longevity risk one is buying them to insure against shows up, inflation is likely to significantly reduce their buying power. 30 years of 2% inflation leaves you with only 55% of the buying power you started with, and 3% inflation would knock that down to only 40%.

This problem can be partly resolved by adding something like a 2-3% fixed COLA, but its usefulness depends on future inflation being relatively close to your COLA. A few years of inflation like we're seeing now could take a big and permanent bite out of a SPIA payout's buying power.
The actual point is that funding almost double the number of years to the age of life expectancy with TIPS is unnecessary. You can fund up to the age of life expectancy with TIPS to reduce inflation risk, and annuitize from the TIPS portfolio at a later age. It is more expensive all else equal to buy a SPIA later because conditional expectation of lifespan conditioned on having attained an older age is longer, so more years are being covered.

But there is a a tendency to overestimate the inflation risk of a SPIA somewhat for various reasons. This has been discussed at length in other threads, from which readers can form their own opinions, so it would not be beneficial to derail this thread with a rehash of that.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

Zeno wrote: Sat Aug 06, 2022 3:37 pm And therein lies the rub. The future is unknowable. Heck, even quantum mechanics is based on statistics and probabilities. If an electron is nothing more than a wave function, I can draw no conclusions regarding the future from looking at the past.
So why invest in index funds?

Which tool used to predict lower future earnings isn't based on history? Certainly not CAPE.
User avatar
willthrill81
Posts: 32250
Joined: Thu Jan 26, 2017 3:17 pm
Location: USA

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by willthrill81 »

Northern Flicker wrote: Sat Aug 06, 2022 4:47 pm
willthrill81 wrote: Sat Aug 06, 2022 9:11 am
Northern Flicker wrote: Sat Aug 06, 2022 1:59 am If the planning horizon is almost double life expectancy, that would make SPIAs very cost effective, as they are priced to fund only up to life expectancy.
The problem with SPIAs these days is that none of them are CPI linked anymore. This means that if the longevity risk one is buying them to insure against shows up, inflation is likely to significantly reduce their buying power. 30 years of 2% inflation leaves you with only 55% of the buying power you started with, and 3% inflation would knock that down to only 40%.

This problem can be partly resolved by adding something like a 2-3% fixed COLA, but its usefulness depends on future inflation being relatively close to your COLA. A few years of inflation like we're seeing now could take a big and permanent bite out of a SPIA payout's buying power.
The actual point is that funding almost double the number of years to the age of life expectancy with TIPS is unnecessary. You can fund up to the age of life expectancy with TIPS to reduce inflation risk, and annuitize from the TIPS portfolio at a later age. It is more expensive all else equal to buy a SPIA later because conditional expectation of lifespan conditioned on having attained an older age is longer, so more years are being covered.

But there is a a tendency to overestimate the inflation risk of a SPIA somewhat for various reasons. This has been discussed at length in other threads, from which readers can form their own opinions, so it would not be beneficial to derail this thread with a rehash of that.
I've never seen a discussion of the "tendency to overestimate the inflation risk of a SPIA" before. That idea makes no sense to me.
nigel_ht
Posts: 4372
Joined: Tue Jan 01, 2019 10:14 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by nigel_ht »

willthrill81 wrote: Sat Aug 06, 2022 4:51 pm
Northern Flicker wrote: Sat Aug 06, 2022 4:47 pm
willthrill81 wrote: Sat Aug 06, 2022 9:11 am
Northern Flicker wrote: Sat Aug 06, 2022 1:59 am If the planning horizon is almost double life expectancy, that would make SPIAs very cost effective, as they are priced to fund only up to life expectancy.
The problem with SPIAs these days is that none of them are CPI linked anymore. This means that if the longevity risk one is buying them to insure against shows up, inflation is likely to significantly reduce their buying power. 30 years of 2% inflation leaves you with only 55% of the buying power you started with, and 3% inflation would knock that down to only 40%.

This problem can be partly resolved by adding something like a 2-3% fixed COLA, but its usefulness depends on future inflation being relatively close to your COLA. A few years of inflation like we're seeing now could take a big and permanent bite out of a SPIA payout's buying power.
The actual point is that funding almost double the number of years to the age of life expectancy with TIPS is unnecessary. You can fund up to the age of life expectancy with TIPS to reduce inflation risk, and annuitize from the TIPS portfolio at a later age. It is more expensive all else equal to buy a SPIA later because conditional expectation of lifespan conditioned on having attained an older age is longer, so more years are being covered.

But there is a a tendency to overestimate the inflation risk of a SPIA somewhat for various reasons. This has been discussed at length in other threads, from which readers can form their own opinions, so it would not be beneficial to derail this thread with a rehash of that.
I've never seen a discussion of the "tendency to overestimate the inflation risk of a SPIA" before. That idea makes no sense to me.
Well, certainly not in 2022...lol.
Zeno
Posts: 769
Joined: Wed Sep 12, 2018 10:44 am

Re: Watch out for hidden risk tolerance assumptions in SWR claims

Post by Zeno »

Great questions
Last edited by Zeno on Sun Aug 07, 2022 8:07 pm, edited 1 time in total.
Post Reply