Retirement planning

BritishInvestor

jamesd said:

BritishInvestor said:

"The UK one for 30 years is 0.3 lower, so 3.7% before costs"

Some studies show ~3.4%

https://finalytiq.co.uk/withdrawal-rates-in-retirement-portfolios-is-the-4-rule-safe-for-uk-clients/

"The effect of costs is a reduction by a third of them, so say take another 0.2 off for 0.6% total costs"

I see it as closer to 0.5%

https://finalytiq.co.uk/impact-of-adviser-fees-on-withdrawal-rates-in-retirement-portfolio/

You've then got to bear in mind typical investor underperformance

https://www.morningstar.com/articles/1056151/why-fund-returns-are-lower-than-you-might-think

I'll ponder that 3.4% result because what I already use is the result of several different sources. You linked to Araham Okuyana's Feb 2015 post and his Nov 2015 post says "More than 100 years of market data for a 60/40 portfolio puts the SWR for the UK at 3.7%" with a 60:40 portfolio. For the moment, given the inherent uncertainties, I remain comfortable with 0.3 lower in the UK for a range of 50-75% equities and 30 years as a decently reasonable one to use. However, I do read what you and others write and ponder whether I should adjust that or whether it's no longer sufficiently good to use just one number. I've been wondering also whether it's worth tying to put together a range of studies of this in one topic to try to consolidate them and provide a handy reference.

I do not agree with the 50% effect of costs in Abraham Okusanya's Feb 2015 post because it seems to be clearly wrong. It used a start date of 1973 which is not the UK  or US worst case. In the worst case the fees are lower because the amount of money on which the fees are charged is depleted faster. It's also somewhat sensitive to duration, being a higher deduction for longer periods, but not to a great enough extent to matter if adjustments of 0.1 or more are being used - which is about the precision limit for any sensible discussion given all of the uncertainties. For around a third or thirty percent I used Kitces analysis at SAFEMAX in the US for 30 years which did have the pot depletion effect. Abraham himself suggests 0.4% in a Guyton-Klinger context in his Jan 2016 comment here: "Because the SWR is the very worse case scenario, every 1% of fees tend to reduce SWR by 0.40%!"

Investors don't seem to be underperforming in that study. It seems to look at the accumulation phase and observe that the gains of early regular investing can be offset by later losses when the pot size is larger, which isn't an investor failure effect. In the decumulation phase it's interesting to consider whether an increase might be seen instead, as the pot size reduces and so will the effect of later decreases in value. However, since SWRs already take cash flows into account it doesn't seem to adjust the SWR value.

"I'll ponder that 3.4% result because what I already use is the result of several different sources. You linked to Araham Okuyana's Feb 2015 post and his Nov 2015 post says "More than 100 years of market data for a 60/40 portfolio puts the SWR for the UK at 3.7%" with a 60:40 portfolio."

I'm assuming Abraham must've refined his data sources over the years - in this article for a 50/50 he gets 3.43% - I now see this as 3.3% in his tool. 

https://finalytiq.co.uk/withdrawal-rates-in-retirement-portfolios-is-the-4-rule-safe-for-uk-clients/

Bumping up to 60/40 sees around a 0.1% improvement.

"I do not agree with the 50% effect of costs in Abraham Okusanya's Feb 2015 post because it seems to be clearly wrong. It used a start date of 1973 which is not the UK  or US worst case"

For my scenarios I get:

1. 50/50, 30 year horizon, no fees: 3.3% (Dec 1968)
2. 50/50, 44 year horizon, no fees: 2.98% (Dec 1968)
3. 50/50, 44 year horizon, 1% fees 2.47% (May 1961)
4. 60/40 44 year horizon, 1% fees, 2.6% (Dec 1968)

Happy to benchmark this against another tool for which we have visibility of the underlying data

"Investors don't seem to be underperforming in that study"

If you look at their comments on dollar-cost averaging, and the relative gap, there definitely seems to be an underperformance - certainly for those that veered away from "boring" investing.

"But investors have fared far better by keeping things simple and sticking with plain-vanilla, broadly diversified funds"

If you look at top-selling funds

https://www.hl.co.uk/funds/research-and-news/popular-funds

there seems to be a bias towards performance chasing/market timing (which doesn't usually end well).

For example, why makes Fundsmith/Baillie Gifford so popular? How many invest in a remotely similar way to the data used for SWR work?

OldScientist

BritishInvestor said:

OldScientist said:

The Pfau study quoted in the finalytiq site uses bills (i.e. something close to cash) rather than bonds to get value of 3.4% for the UK, Pfau's earlier work (using bonds, https://www.advisorperspectives.com/articles/2014/03/04/does-international-diversification-improve-safe-withdrawal-rates) gave a UK safemax of 3.1% for a 50/50 portfolio. Estrada (Maximum Withdrawal rates: An empirical and Global Perspective, 2016) has, as far as I am aware, the most comprehensive published analysis of UK withdrawal rates (and the performance of many other countries too) with safe withdrawal rates (at 1% failure - values will be close enough to Safemax) ranging from 3.1% (portfolio 50/50 stocks/bonds) to 3.5% (at 80/20) - for portfolios with plenty of fixed income, using bills instead of bonds might raise these. According to the values presented by Estrada, the difference between UK and USA SWR is around 0.5% (the exact value depends on the asset allocation).

FWIW I get 3.3% using 50/50 global equities and global agg bonds, so as you point out, underlying data can make a difference.

With apologies to the OP for the off-topic question. Definitely true - can I ask what data set you're using? The only comprehensive free one I've found is the JST database (https://www.macrohistory.net).

OldScientist

BritishInvestor said:

jamesd said:

BritishInvestor said:

"The UK one for 30 years is 0.3 lower, so 3.7% before costs"

Some studies show ~3.4%

https://finalytiq.co.uk/withdrawal-rates-in-retirement-portfolios-is-the-4-rule-safe-for-uk-clients/

"The effect of costs is a reduction by a third of them, so say take another 0.2 off for 0.6% total costs"

I see it as closer to 0.5%

https://finalytiq.co.uk/impact-of-adviser-fees-on-withdrawal-rates-in-retirement-portfolio/

You've then got to bear in mind typical investor underperformance

https://www.morningstar.com/articles/1056151/why-fund-returns-are-lower-than-you-might-think

I'll ponder that 3.4% result because what I already use is the result of several different sources. You linked to Araham Okuyana's Feb 2015 post and his Nov 2015 post says "More than 100 years of market data for a 60/40 portfolio puts the SWR for the UK at 3.7%" with a 60:40 portfolio. For the moment, given the inherent uncertainties, I remain comfortable with 0.3 lower in the UK for a range of 50-75% equities and 30 years as a decently reasonable one to use. However, I do read what you and others write and ponder whether I should adjust that or whether it's no longer sufficiently good to use just one number. I've been wondering also whether it's worth tying to put together a range of studies of this in one topic to try to consolidate them and provide a handy reference.

I do not agree with the 50% effect of costs in Abraham Okusanya's Feb 2015 post because it seems to be clearly wrong. It used a start date of 1973 which is not the UK  or US worst case. In the worst case the fees are lower because the amount of money on which the fees are charged is depleted faster. It's also somewhat sensitive to duration, being a higher deduction for longer periods, but not to a great enough extent to matter if adjustments of 0.1 or more are being used - which is about the precision limit for any sensible discussion given all of the uncertainties. For around a third or thirty percent I used Kitces analysis at SAFEMAX in the US for 30 years which did have the pot depletion effect. Abraham himself suggests 0.4% in a Guyton-Klinger context in his Jan 2016 comment here: "Because the SWR is the very worse case scenario, every 1% of fees tend to reduce SWR by 0.40%!"

Investors don't seem to be underperforming in that study. It seems to look at the accumulation phase and observe that the gains of early regular investing can be offset by later losses when the pot size is larger, which isn't an investor failure effect. In the decumulation phase it's interesting to consider whether an increase might be seen instead, as the pot size reduces and so will the effect of later decreases in value. However, since SWRs already take cash flows into account it doesn't seem to adjust the SWR value.

"I'll ponder that 3.4% result because what I already use is the result of several different sources. You linked to Araham Okuyana's Feb 2015 post and his Nov 2015 post says "More than 100 years of market data for a 60/40 portfolio puts the SWR for the UK at 3.7%" with a 60:40 portfolio."

I'm assuming Abraham must've refined his data sources over the years - in this article for a 50/50 he gets 3.43% - I now see this as 3.3% in his tool. 

https://finalytiq.co.uk/withdrawal-rates-in-retirement-portfolios-is-the-4-rule-safe-for-uk-clients/

Bumping up to 60/40 sees around a 0.1% improvement.

The 3.7% in the November post appears to come from the Kitces presentation in late 2015 (https://finalytiq.co.uk/wp-content/uploads/2012/08/Sequence-Of-Return-Risk-Safe-Withdrawal-Rates-Retirement-Income-Conference-w-Aviva-FinalytiQ-Handouts.pdf) - the key graph is on page 11 - 3.7% is with a 90/10 portfolio (with bonds, it is slightly higher with bills) - at 50/50 the values are more like 3.1 and 3.3% respectively. I don't have Okuyana's book, but there is a link to a graph from it (at least that is the reference) at https://moneyunshackled.com/2021/01/triple-your-pension-income-optimal-safe-withdrawal-rate/ which suggests that the UK SWR is 3.1% (i.e. the same as Pfau and Estrada if the portfolio in that graph is 50/50).

BritishInvestor

OldScientist said:

BritishInvestor said:

jamesd said:

BritishInvestor said:

"The UK one for 30 years is 0.3 lower, so 3.7% before costs"

Some studies show ~3.4%

https://finalytiq.co.uk/withdrawal-rates-in-retirement-portfolios-is-the-4-rule-safe-for-uk-clients/

"The effect of costs is a reduction by a third of them, so say take another 0.2 off for 0.6% total costs"

I see it as closer to 0.5%

https://finalytiq.co.uk/impact-of-adviser-fees-on-withdrawal-rates-in-retirement-portfolio/

You've then got to bear in mind typical investor underperformance

https://www.morningstar.com/articles/1056151/why-fund-returns-are-lower-than-you-might-think

I'll ponder that 3.4% result because what I already use is the result of several different sources. You linked to Araham Okuyana's Feb 2015 post and his Nov 2015 post says "More than 100 years of market data for a 60/40 portfolio puts the SWR for the UK at 3.7%" with a 60:40 portfolio. For the moment, given the inherent uncertainties, I remain comfortable with 0.3 lower in the UK for a range of 50-75% equities and 30 years as a decently reasonable one to use. However, I do read what you and others write and ponder whether I should adjust that or whether it's no longer sufficiently good to use just one number. I've been wondering also whether it's worth tying to put together a range of studies of this in one topic to try to consolidate them and provide a handy reference.

I do not agree with the 50% effect of costs in Abraham Okusanya's Feb 2015 post because it seems to be clearly wrong. It used a start date of 1973 which is not the UK  or US worst case. In the worst case the fees are lower because the amount of money on which the fees are charged is depleted faster. It's also somewhat sensitive to duration, being a higher deduction for longer periods, but not to a great enough extent to matter if adjustments of 0.1 or more are being used - which is about the precision limit for any sensible discussion given all of the uncertainties. For around a third or thirty percent I used Kitces analysis at SAFEMAX in the US for 30 years which did have the pot depletion effect. Abraham himself suggests 0.4% in a Guyton-Klinger context in his Jan 2016 comment here: "Because the SWR is the very worse case scenario, every 1% of fees tend to reduce SWR by 0.40%!"

Investors don't seem to be underperforming in that study. It seems to look at the accumulation phase and observe that the gains of early regular investing can be offset by later losses when the pot size is larger, which isn't an investor failure effect. In the decumulation phase it's interesting to consider whether an increase might be seen instead, as the pot size reduces and so will the effect of later decreases in value. However, since SWRs already take cash flows into account it doesn't seem to adjust the SWR value.

"I'll ponder that 3.4% result because what I already use is the result of several different sources. You linked to Araham Okuyana's Feb 2015 post and his Nov 2015 post says "More than 100 years of market data for a 60/40 portfolio puts the SWR for the UK at 3.7%" with a 60:40 portfolio."

I'm assuming Abraham must've refined his data sources over the years - in this article for a 50/50 he gets 3.43% - I now see this as 3.3% in his tool. 

https://finalytiq.co.uk/withdrawal-rates-in-retirement-portfolios-is-the-4-rule-safe-for-uk-clients/

Bumping up to 60/40 sees around a 0.1% improvement.

The 3.7% in the November post appears to come from the Kitces presentation in late 2015 (https://finalytiq.co.uk/wp-content/uploads/2012/08/Sequence-Of-Return-Risk-Safe-Withdrawal-Rates-Retirement-Income-Conference-w-Aviva-FinalytiQ-Handouts.pdf) - the key graph is on page 11 - 3.7% is with a 90/10 portfolio (with bonds, it is slightly higher with bills) - at 50/50 the values are more like 3.1 and 3.3% respectively. I don't have Okuyana's book, but there is a link to a graph from it (at least that is the reference) at https://moneyunshackled.com/2021/01/triple-your-pension-income-optimal-safe-withdrawal-rate/ which suggests that the UK SWR is 3.1% (i.e. the same as Pfau and Estrada if the portfolio in that graph is 50/50).

Right, I think I see where the differences lie.

Looking at that chart in the book the SWR is:

1. using UK equities and bonds
2. based on a starting point of 1900

Timeline only goes back to 1915, and I use global equities and bonds (but still UK inflation) which would explain my 3.3% (global equities and bonds) vs your 3.1%. It does go to show that ~100 years of data really isn't that long, so the SWR cannot be 100% relied upon. 

FWIW if I put in UK assets I get 3.35% (July 1947) when using UK agg bonds.

BritishInvestor

OldScientist said:

BritishInvestor said:

OldScientist said:

The Pfau study quoted in the finalytiq site uses bills (i.e. something close to cash) rather than bonds to get value of 3.4% for the UK, Pfau's earlier work (using bonds, https://www.advisorperspectives.com/articles/2014/03/04/does-international-diversification-improve-safe-withdrawal-rates) gave a UK safemax of 3.1% for a 50/50 portfolio. Estrada (Maximum Withdrawal rates: An empirical and Global Perspective, 2016) has, as far as I am aware, the most comprehensive published analysis of UK withdrawal rates (and the performance of many other countries too) with safe withdrawal rates (at 1% failure - values will be close enough to Safemax) ranging from 3.1% (portfolio 50/50 stocks/bonds) to 3.5% (at 80/20) - for portfolios with plenty of fixed income, using bills instead of bonds might raise these. According to the values presented by Estrada, the difference between UK and USA SWR is around 0.5% (the exact value depends on the asset allocation).

FWIW I get 3.3% using 50/50 global equities and global agg bonds, so as you point out, underlying data can make a difference.

With apologies to the OP for the off-topic question. Definitely true - can I ask what data set you're using? The only comprehensive free one I've found is the JST database (https://www.macrohistory.net).

I'm using https://www.timelineapp.co/

Audaxer

MK62 said:

The Which website estimates £41k for a luxury retirement lifestyle for a couple, and I'm sure I saw a link from this forum to another site recently which had £47k for a luxury retirement lifestyle for a couple.

In the end this is going to come down to personal opinion on the meaning of luxury.......personally I don't see £41k for a couple as luxury....perhaps not even £47k either........on the sunny side of comfortable, sure, but luxury???

I agree, but the point I was making is that these figures are much more realistic than the quoted figure of £20,978 for a 5* retirement lifestyle.

jamesd

The Which survey of members and estimated spending is here. The Retirement Living Standards are here. Each describes what spending it is based on at each level.

MK62

Audaxer said:

MK62 said:

The Which website estimates £41k for a luxury retirement lifestyle for a couple, and I'm sure I saw a link from this forum to another site recently which had £47k for a luxury retirement lifestyle for a couple.

In the end this is going to come down to personal opinion on the meaning of luxury.......personally I don't see £41k for a couple as luxury....perhaps not even £47k either........on the sunny side of comfortable, sure, but luxury???

I agree, but the point I was making is that these figures are much more realistic than the quoted figure of £20,978 for a 5* retirement lifestyle.

Totally agree.....for us £20978 would be in the breadline/penury category.....perhaps they mean this sum in addition to state pensions......that would make more sense (but still short of "luxury" imho)...

flopsy1973

We don't have a very expensive lifestyle hence the figures quoted the only extra maybe would be travelling which I am keen to do but that would be in region of 3k a year maybe more some years. I am just worried  about the gap until state pension  and then  we will need lot less. Her parents do it living in Spain on less than figures above and possibly we will move out there one day. So could we o it ? 

Albermarle

MK62 said:

Audaxer said:

MK62 said:

The Which website estimates £41k for a luxury retirement lifestyle for a couple, and I'm sure I saw a link from this forum to another site recently which had £47k for a luxury retirement lifestyle for a couple.

In the end this is going to come down to personal opinion on the meaning of luxury.......personally I don't see £41k for a couple as luxury....perhaps not even £47k either........on the sunny side of comfortable, sure, but luxury???

I agree, but the point I was making is that these figures are much more realistic than the quoted figure of £20,978 for a 5* retirement lifestyle.

Totally agree.....for us £20978 would be in the breadline/penury category.....perhaps they mean this sum in addition to state pensions......that would make more sense (but still short of "luxury" imho)...

As said previously , I do not think you could even live a proper 5*/luxury lifestyle on the figures often quoted ( about £45K for a couple after tax ) If you are going to drive around in top of the range Range Rovers, send kids to private school, First class cruises, long haul holidays in Business class, expensive sports/hobbies , big house with regular top end refurbishments , busy social whirl etc etc , you would need at least double that .