Pension Drawdown Calculators

dunstonh

Sorry you've lost me a bit - is your graph above showing that the Which calculation is wrong, or are you saying that your graph is  based on real history from 2000 to 2009 and showing that if I had retired in 2000, this is not what would have happened?

The first graph is not a projection of what may happen.  It is the real 10 year period between 2000-2009.

However the Which model doesn't claim that the fund is 100% in stocks and of course you have picked the period that shows a massive collapse in value early in the retirement.

You never know when losses are going to happen.  They could happen in year one.  They may not.   You know they are going to happen though.  The last 14 years have been perfect for investing.  It has resulted in higher returns than most periods.   Whilst there is no pattern, you do often find that periods alternate between good and bad.  Partly as the things that create good periods have to go in the opposite direction at some point.  For example, fixed interest securities do well when interest rates fall and inflation falls.  They do badly when either of those rise or you get the perfect storm of both of them rising.

The example model you looked at showed equities growing at a level return of 7% a year.   Yet you can see from that example, that isn't how it works.

And for reference, here is a multi-asset example on the same basis as the post above. It models worse, not better.

The Which model allowed you to pick an inflation value rather then forecasting - I picked 2%.

Picking an artificially low rate that has only happened in a minority of periods in the UK is not a good way to plan for 30-40 years.  Inflation could hit 10% this year.

Of course if I picked 5% it would look more like your second chart above. 

But 5% is closer to the long term average for the UK.   So, why do you think that 2% is better than 5%?

However keeping in mind that I am not obliged to increase my drawdown by inflation every year - I could rather cut discretionary spending if needed.

Whereas with the annually increasing DB pension, you wouldn't need to.  

Again I am not claiming this would be my good decision - just trying to understand all the angles on it.

Relying on optimistic assumptions is going to hurt sooner or later.  

Which is a pretty long standing reputable organisation who I think have been running a money magazine and service for many years, so I would hope that anything they put out is not total garbage - at least not based on history. 

You are old enough to remember the years they recommended endowments.  And the method they recommended purchasing them resulted in people being unable to complain and get compensation.  To be fair they couldn't predict what would happen but you should not be reliant on modelling examples that use straight line figures 

It is impossible to model the future as you do not know what the future holds.  You can model what the past would have done, which is probably one of the better ways.  However, too many models use very simplistic fixed rates.

The important thing is not to be optimistic and use projections that may have worked in the last decade but wont work in all.

Pat38493

dunstonh said:

Sorry you've lost me a bit - is your graph above showing that the Which calculation is wrong, or are you saying that your graph is  based on real history from 2000 to 2009 and showing that if I had retired in 2000, this is not what would have happened?

The first graph is not a projection of what may happen.  It is the real 10 year period between 2000-2009.

However the Which model doesn't claim that the fund is 100% in stocks and of course you have picked the period that shows a massive collapse in value early in the retirement.

You never know when losses are going to happen.  They could happen in year one.  They may not.   You know they are going to happen though.  The last 14 years have been perfect for investing.  It has resulted in higher returns than most periods.   Whilst there is no pattern, you do often find that periods alternate between good and bad.  Partly as the things that create good periods have to go in the opposite direction at some point.  For example, fixed interest securities do well when interest rates fall and inflation falls.  They do badly when either of those rise or you get the perfect storm of both of them rising.

The example model you looked at showed equities growing at a level return of 7% a year.   Yet you can see from that example, that isn't how it works.

And for reference, here is a multi-asset example on the same basis as the post above. It models worse, not better.

The Which model allowed you to pick an inflation value rather then forecasting - I picked 2%.

Picking an artificially low rate that has only happened in a minority of periods in the UK is not a good way to plan for 30-40 years.  Inflation could hit 10% this year.

Of course if I picked 5% it would look more like your second chart above. 

But 5% is closer to the long term average for the UK.   So, why do you think that 2% is better than 5%?

However keeping in mind that I am not obliged to increase my drawdown by inflation every year - I could rather cut discretionary spending if needed.

Whereas with the annually increasing DB pension, you wouldn't need to.  

Again I am not claiming this would be my good decision - just trying to understand all the angles on it.

Relying on optimistic assumptions is going to hurt sooner or later.  

Which is a pretty long standing reputable organisation who I think have been running a money magazine and service for many years, so I would hope that anything they put out is not total garbage - at least not based on history. 

You are old enough to remember the years they recommended endowments.  And the method they recommended purchasing them resulted in people being unable to complain and get compensation.  To be fair they couldn't predict what would happen but you should not be reliant on modelling examples that use straight line figures 

It is impossible to model the future as you do not know what the future holds.  You can model what the past would have done, which is probably one of the better ways.  However, too many models use very simplistic fixed rates.

The important thing is not to be optimistic and use projections that may have worked in the last decade but wont work in all.

Yes I guess my inflation number is way too optimistic as I was probably looking at historical inflation in recent memory of the last decade or two.

The other model that I quoted above based on the last 120 years or so, claims that only 2.5% of the time, so 3 years out of the last 120 years, would you have run out of money in 18 years based on that scenario, so whilst I picked way to optimistic, I wonder if you have picked the worst case year to measure it on.  Or maybe that other model is wrong, but that model claims to have incorporated the real historical inflation into all its calculations as well.

dunstonh

whilst I picked way to optimistic, I wonder if you have picked the worst case year to measure it on.  Or maybe that other model is wrong, but that model claims to have incorporated the real historical inflation into all its calculations as well.

Below is a £750k with values adjusted for inflation and starting at age 55 with the maximum amount you could draw each year without the money running about before age 98 (it assumes £1 left in the post at age 98).  Green indicates the best period, red the worst.  The purple line shows your £31,500 draw



It indicates that only in 26% of time periods a £31,500 draw would have been sustainable without running out of money.  That is based on a 50/50 mixed asset portfolio. 
 So, apart from the post WWI boom the period of globalisation in recent times the rest of the periods failed or only just made it.  Globalisation is now in decline.   Commodities are heavily controlled by China and Russia.  So, does that bode well for pushing limits?

The same charting but assuming age 87 as the point of running out:
That is a 52% success rate.

Pat38493

dunstonh said:

whilst I picked way to optimistic, I wonder if you have picked the worst case year to measure it on.  Or maybe that other model is wrong, but that model claims to have incorporated the real historical inflation into all its calculations as well.

Below is a £750k with values adjusted for inflation and starting at age 55 with the maximum amount you could draw each year without the money running about before age 98 (it assumes £1 left in the post at age 98).  Green indicates the best period, red the worst.  The purple line shows your £31,500 draw



It indicates that only in 26% of time periods a £31,500 draw would have been sustainable without running out of money.  That is based on a 50/50 mixed asset portfolio. 
 So, apart from the post WWI boom the period of globalisation in recent times the rest of the periods failed or only just made it.  Globalisation is now in decline.   Commodities are heavily controlled by China and Russia.  So, does that bode well for pushing limits?

The same charting but assuming age 87 as the point of running out:
That is a 52% success rate.

Very interesting.  The other model I mentioned shows 44% compared to your model 26% based on the same numbers plugged in, and they don’t give enough information to analyse why your model is significantly different to theirs as the high level claims seems the same.

Actually - the other model claims to be looking at 121 years of data, but, it does not make sense to include data for times which goes outside the period you are requesting and would require forecasting - e.g. according to their description, if I am looking over 43 years, they should only be looking at retirement dates up to 1979.  Maybe this is automatically done behind the scenes but it doesn’t specifically say that.

Thrugelmir

A link for you. Unfortunately only summaries of the data can be accessed for free. 

https://www.credit-suisse.com/about-us-news/en/articles/media-releases/credit-suisse-global-investment-returns-yearbook-2022-202202.html

Pat38493

Thrugelmir said:

A link for you. Unfortunately only summaries of the data can be accessed for free. 

https://www.credit-suisse.com/about-us-news/en/articles/media-releases/credit-suisse-global-investment-returns-yearbook-2022-202202.html

Thanks I will take a look.

It’s worth mentioning that although this thread was based on the probably preposterous notion that if I cashed in my DB, I could retire at 55 and take the full non discounted amount out, and still live beyond the normal lifespan, if you draw a line on dunstonh’s chart at £18.3K, I think the line is always below every bar. 

£18.3K is the number that was quoted to me by the DB scheme as an estimated retirement income age 57

As such, if I was only deciding based on this historical data, I could conclude that in no year, would I have been better off by keeping the DB scheme in place (unless of course I live beyond 98 and still need the same drawdown).

Of course as pointed out, this doesn’t necessarily mean you should exchange a cast iron guarantee for a high probability, but it’s nevertheless quite interesting, in the context that there are there are many years where you could be substantially better off than the £18.3K.

(by the way my DB CETV value was 771K not 750K so it was a bit higher than I remembered).

Thrugelmir

In the same way that night follows day. Always a new generation of investors that on the back of a long bull market that believe that " This Time Is Different".  Said to be the 4 more expensive words in the English language. Confirmation bias abounds.  High CETV's are offered for good reason. Pension schemes want to offload liabilities while markets are exceptionally buoyant. Once markets fall back CETV's will do likewise. 

Pat38493

Thrugelmir said:

In the same way that night follows day. Always a new generation of investors that on the back of a long bull market that believe that " This Time Is Different".  Said to be the 4 more expensive words in the English language. Confirmation bias abounds.  High CETV's are offered for good reason. Pension schemes want to offload liabilities while markets are exceptionally buoyant. Once markets fall back CETV's will do likewise. 

Yes this is actually the thing I am also thinking now.  If pension companies are offering CETV values that seem to be better than what they would have needed to fund any liability in the last more than 100 years, the question is why.

Either
- They think I will live beyond 98 regardless of current general life expectancy.
- They believe the next few years will be the worst time to retire for more than 100 years in terms of pension pot management.
- They want offload other potential unknown risks like death benefits, dependent pensions etc.

However - doesn’t this mean that if I had the other means to avoid making any withdrawals from the capital in, say, the next 10 years, it could actually be a good deal to take the current CETV?

Also just to point out that the historical data we are quoting goes back over 100 years so it is not only relating to recent good performances.

OldScientist

Pat38493 said:

It states
"That’s assuming that once you retire, your cash investment grows at an average of 0.50% a year, fixed interest at 4.75% a year and equities at 7.25% a year."

[snip]

I take your point about inflation but I am not retiring this year and historically inflation average seems to have been about 2%, but probably you are right that this is an "optimistic" view.

All that said, Which is a pretty long standing reputable organisation who I think have been running a money magazine and service for many years, so I would hope that anything they put out is not total garbage - at least not based on history.  Of course if the future is massively different from the past no model will be correct.

The first statement is why there is a significant difference between the Which and 2020finance calculators - for the Which calculator, taking average returns gives average/median results - in reality, different sequences of returns means that some historical retirements will do better and others worse (a lot worse).

Average inflation is highly period dependent - the UK average inflation from 1989 to now is about 2.5%, project back just a few more years to 1960 and the average inflation is over 5%.

OldScientist

Pat38493 said:

dunstonh said:

whilst I picked way to optimistic, I wonder if you have picked the worst case year to measure it on.  Or maybe that other model is wrong, but that model claims to have incorporated the real historical inflation into all its calculations as well.

Below is a £750k with values adjusted for inflation and starting at age 55 with the maximum amount you could draw each year without the money running about before age 98 (it assumes £1 left in the post at age 98).  Green indicates the best period, red the worst.  The purple line shows your £31,500 draw



It indicates that only in 26% of time periods a £31,500 draw would have been sustainable without running out of money.  That is based on a 50/50 mixed asset portfolio. 
 So, apart from the post WWI boom the period of globalisation in recent times the rest of the periods failed or only just made it.  Globalisation is now in decline.   Commodities are heavily controlled by China and Russia.  So, does that bode well for pushing limits?

The same charting but assuming age 87 as the point of running out:
That is a 52% success rate.

Very interesting.  The other model I mentioned shows 44% compared to your model 26% based on the same numbers plugged in, and they don’t give enough information to analyse why your model is significantly different to theirs as the high level claims seems the same.

Actually - the other model claims to be looking at 121 years of data, but, it does not make sense to include data for times which goes outside the period you are requesting and would require forecasting - e.g. according to their description, if I am looking over 43 years, they should only be looking at retirement dates up to 1979.  Maybe this is automatically done behind the scenes but it doesn’t specifically say that.

The 2020finance site uses the Barclay's dataset of returns and inflation (there's some information towards the bottom) which is a well regarded dataset, although there are arguably better ones. The inflation values are quite optimistic compared to the ones held by the ONS (prior to 1949 there was no single official measure of inflation in the UK and there are at least 4 different attempts to estimate historical inflation).

I note 2020finance say that "in simulations that go beyond the present year, it will wrap back to 1928 and count up from there" which may explain why there are 122 simulations regardless of what duration you set.

Dunstonh - please satisfy my curiosity - what package are you using there (and are the years on the x-axis start years)?