Growth assumptions in your models/spreadsheets

NoMore

Hoenir said:

NoMore said:

Hoenir said:

westv said:

Hoenir said:

QrizB said:

michaels said:

2) I certainly think there is room for an annuity when rates exceed historic SWR (or whatever other 'worst case return scenario you use in your modelling)...

The UK SWR is something like 3.5% but a single-life RPI annuity is yielding something like 4.8% for a 65-year-old retiree.

SWR isn't guaranted to track inflation though. There's no direct correlation between equity returns and inflation over any given period of time. . 

The whole point of the SWR is to provide an income that matches inflation.

SWR would have to have rules applied in that case. Drawn income levels would have to be flexible. 

You don't appear to understand how the SWR works. Its not a % amount of the remaining pot each year. That would result in a variable income dependant on investment return. Its at the start you take the SWR of your initial pot and then increase that £ amount each year by inflation. The initial percentage of pot is only used at the start. Its designed to give you the same income, increasing by inflation each year, no matter what investment returns will be, based on historic returns.

Unsurprisingly this topc is an old chesnut.  Where the same fallacies arise time over time.  Be wonderfull if (a) investments did in fact provide an inflation linked return (b) there was a direct correlation between inflation and investment returns. The rules that accompanied the US research on SWR are well documented. 

I don't understand what you are saying here. Are you saying that I am wrong about how the SWR works in respect to inflation ? Yes investments don't provide a guaranteed inflation linked return but the SWR isn't about that, its looking back at history and determining what would have be the highest amount adjusted for inflation each year you could have withdrawn and not run out of money over a expected period. Of course future returns are unknown so doesn't guarantee you won't run out of money in the future.

westv

NoMore said:

Hoenir said:

NoMore said:

Hoenir said:

westv said:

Hoenir said:

QrizB said:

michaels said:

2) I certainly think there is room for an annuity when rates exceed historic SWR (or whatever other 'worst case return scenario you use in your modelling)...

The UK SWR is something like 3.5% but a single-life RPI annuity is yielding something like 4.8% for a 65-year-old retiree.

SWR isn't guaranted to track inflation though. There's no direct correlation between equity returns and inflation over any given period of time. . 

The whole point of the SWR is to provide an income that matches inflation.

SWR would have to have rules applied in that case. Drawn income levels would have to be flexible. 

You don't appear to understand how the SWR works. Its not a % amount of the remaining pot each year. That would result in a variable income dependant on investment return. Its at the start you take the SWR of your initial pot and then increase that £ amount each year by inflation. The initial percentage of pot is only used at the start. Its designed to give you the same income, increasing by inflation each year, no matter what investment returns will be, based on historic returns.

Unsurprisingly this topc is an old chesnut.  Where the same fallacies arise time over time.  Be wonderfull if (a) investments did in fact provide an inflation linked return (b) there was a direct correlation between inflation and investment returns. The rules that accompanied the US research on SWR are well documented. 

I don't understand what you are saying here. Are you saying that I am wrong about how the SWR works in respect to inflation ? Yes investments don't provide a guaranteed inflation linked return but the SWR isn't about that, its looking back at history and determining what would have be the highest amount adjusted for inflation each year you could have withdrawn and not run out of money over a expected period. Of course future returns are unknown so doesn't guarantee you won't run out of money in the future.

I am confused by the response too. I don't think they have really understood the point being made perhaps?

OldScientist

MK62 said:

OldScientist said:

michaels said:

SimonSeys said:

NoMore said:

Hoenir said:

westv said:

Hoenir said:

QrizB said:

michaels said:

2) I certainly think there is room for an annuity when rates exceed historic SWR (or whatever other 'worst case return scenario you use in your modelling)...

The UK SWR is something like 3.5% but a single-life RPI annuity is yielding something like 4.8% for a 65-year-old retiree.

SWR isn't guaranted to track inflation though. There's no direct correlation between equity returns and inflation over any given period of time. . 

The whole point of the SWR is to provide an income that matches inflation.

SWR would have to have rules applied in that case. Drawn income levels would have to be flexible. 

You don't appear to understand how the SWR works. Its not a % amount of the remaining pot each year. That would result in a variable income dependant on investment return. Its at the start you take the SWR of your initial pot and then increase that £ amount each year by inflation. The initial percentage of pot is only used at the start. Its designed to give you the same income, increasing by inflation each year, no matter what investment returns will be, based on historic returns.

The reason people don't tend to follow it religiously, is it could take a lot of nerve to continue to take what could be a large percentage of your pot when the investment is not performing. 

I'm not advocating it as a strategy, just explaining how it actually works.

Doesn’t that mean that SWR is time dependent?  I’d the SWR for someone aged 50 is different to someone aged 80

Yes, an SWR is a real terms annual amount (expressed as a proportion of the initial pot) for 'zero historic failures' over a given number of years.  Generally 30 years is quoted although of course many of us would be looking at non-zero odds of a 40 year plus retirement - however as the number of years increases the SWR tends to converge to a fixed value (about 3%) rather than continuing to decline.

Just to add some data...

Using the calculator at https://www.2020financial.co.uk/pension-drawdown-calculator/ with a 60/20/20 (Uk equities/long bonds/cash) portfolio the safemax (i.e., the SWR where no failures occurred) was

Expected length of retirement at retirement

50    40    30    20

2.6   2.8   3.3   4.1

Assuming a longevity to 100yo (for a couple there is about a 10% chance of one or other or both living that long) then these would be for a retirement at ages 50, 60, 70, or 80.

Interestingly, the calculator also says the safemax at 60/40 equities/cash (ie no bonds) was 

Retirement length
50       40      30      20

2.56   3.11   3.66   4.51

For comparison, 100% cash shows a 30y safemax of 2.82%, and 100% equities, 3.43% and 100% bonds, 1.49%

While the calculator is a useful tool (and, AFAIK, the currently the only free one using historical data for UK retirements), it is definitely worth looking at the assumptions tab towards the end of the calculator.

Bond returns are from Barclays equity gilt study where despite what the calculator says, before about 1962 they were returns on consols, from 1962 to 1980 or so, returns on 20 year gilts, and after 1980 15 year gilts. In other words, very long duration even compared to the 'all stocks' fund. Intermediate duration funds would have done better.

I think their model will underestimate the returns on cash when rates rise and overestimate it when rates fall. Certainly using the returns at macrohistory.net (which uses the same source for bond returns) gets a safemax of 1.5% for all bonds, 1.9% for all cash (actually 3-month bills) and 3.3% for all equities - i.e., close enough except for cash.

Peterrr

Thanks OldScientist. Am I correct in understanding from those safemax rates that cash has provided a better average return than bonds? 

MK62

Peterrr said:

Thanks OldScientist. Am I correct in understanding from those safemax rates that cash has provided a better average return than bonds? 

No......the safemax rates given by the calculator are telling you that bonds (actually in this case UK gilts) had a worse "worst case" than cash......they are not telling you anything about average returns......however, the calculator does touch on that with some info about minimum, maximum and average final pot size.

OldScientist

MK62 said:

Peterrr said:

Thanks OldScientist. Am I correct in understanding from those safemax rates that cash has provided a better average return than bonds? 

No......the safemax rates given by the calculator are telling you that bonds (actually in this case UK gilts) had a worse "worst case" than cash......they are not telling you anything about average returns......however, the calculator does touch on that with some info about minimum, maximum and average final pot size.

The 'bonds' in the online calculator are UK long gilts (15 year maturities or greater), 'all stocks' or intermediate duration funds (e.g., under 10 year) would have done better.

On a quick look, I've found that the difference in performance for 'cash' in the calculator I linked to and 'cash' in the calculator I've developed using the macrohistory.net data is the instruments used to represent cash (and the inflation model). The online calculator uses the bank rate (what used to be called the minimum lending rate in the 70s and dealing rate in the 80s and 90s), while macrohistory.net uses the returns from 3-month treasury bills. In one of the worst cases (a retirement starting in 1935) the bank rate was 2% until 1951 while the bill rate lay between 0.5 and 1%, and this makes a significant difference to the outcome (as does the inflation model).

MK62

Also, the calculator uses UK gilts for the "bonds" no corporate, no overseas, no high yield etc..........and uses UK equities for the "stocks".......but I suspect few these days invest their entire equity holding in just UK equities (though they might have years ago).

Hoenir

MK62 said:

Also, the calculator uses UK gilts for the "bonds" no corporate, no overseas, no high yield etc..........and uses UK equities for the "stocks".......but I suspect few these days invest their entire equity holding in just UK equities (though they might have years ago).

How would you factor exchange rates into projecting future returns as well? The permutations are endless. 

MK62

Hoenir said:

MK62 said:

Also, the calculator uses UK gilts for the "bonds" no corporate, no overseas, no high yield etc..........and uses UK equities for the "stocks".......but I suspect few these days invest their entire equity holding in just UK equities (though they might have years ago).

How would you factor exchange rates into projecting future returns as well? The permutations are endless. 

I didn't say I would......projecting future returns is always going to be a bit of a guessing game really - I accept that the only "data" we have is the historic stuff, so there's really not much choice there ....but I agree that exchange rates would add significantly more complexity to the equation.

OldScientist

Hoenir said:

MK62 said:

Also, the calculator uses UK gilts for the "bonds" no corporate, no overseas, no high yield etc..........and uses UK equities for the "stocks".......but I suspect few these days invest their entire equity holding in just UK equities (though they might have years ago).

How would you factor exchange rates into projecting future returns as well? The permutations are endless. 

You're right. Even getting a handle on the historical SWR depends, as might be expected, on a lot of things: overall asset allocation, composition of assets (e.g., GDP weighted, cap weighted, domestic or international, duration of  fixed income etc.), and even the inflation model (e.g., there are at least four pre-1948 datasets of UK inflation - that contribute to a 20-40 bp difference in SWR depending on asset allocation). Uncertainties in the historical SWR have not been much addressed, but getting it pinned down to the nearest 50 bp is probably good going (using 2 decimal places is definitely going well beyond the actual precision). In other words, the rule of thumb often used round here of somewhere between 3.0% and 3.5% for 30 years is good enough.

I think the point is that future growth is unknown, future SWR is unknown, and (for adaptive withdrawal strategies) future income is unknown and that uncertainty needs to be incorporated into planning - my own preference has been to establish a floor of guaranteed income (not without its own risks, but that is a different conversation) that covers core expenditure and much of our discretionary needs and therefore we don't much care about variations in portfolio withdrawals. Such an approach will not suit everyone.