SWR Question

ac198179

Hi,

I've got a question about the SWR in retirement. Assume a 4% SWR, I thought that each year you took out 4% of your pot (plus adjustment each year for inflation, but let's ignore inflation just for now). But is it actually that each you take out 4% of your starting balance at retirement? E.g. say I retire at 57 with a pot of £100k, 4% of which is of course £4k - is that how much I withdraw on each of the following years, or do I withdraw 4% of the new balance each year? So, assuming no growth at all between year 0 and year 1, do I then withdraw 4% of £96k?

Thanks!

taking_stock

It's 4% of the original sum, so £4k in the first year and the same amount plus inflation thereafter.

[Deleted User]

That’s how SWR is designed but it does not actually exist.  By definition you cannot have a “safe” constant rate of withdrawal of a highly variable market dependent position.  

Croeso69

Deleted_User said:

That’s how SWR is designed but it does not actually exist.  By definition you cannot have a “safe” constant rate of withdrawal of a highly variable market dependent position.  

A SWR does exist. I am pretty sure a rate of 1% could be supported but you would need a big fund to support a worthwhile amount to live on.

JohnWinder

Using your retire at 57 example, you'd withdraw £4k in the first year, in the second year you would withdraw a pound amount which was £4k plus however much the inflation rate had increased £4k by, so for 10% inflation you'd withdraw £4400.

Truely, you can't have a perfectly safe constant rate of withdrawal of this nature from a pot that can fluctuate any old way, but imagine withdrawing 0.01%/year (instead of 4%) as your SWR using the method described, and tell me whether you'd be worried about ever running out of money. There's theory, and there's being practical.

https://www.bogleheads.org/wiki/Safe_withdrawal_rates

ac198179

Thanks very much everyone! It's a lot clearer now!

[Deleted User]

Croeso69 said:

Deleted_User said:

That’s how SWR is designed but it does not actually exist.  By definition you cannot have a “safe” constant rate of withdrawal of a highly variable market dependent position.  

A SWR does exist. I am pretty sure a rate of 1% could be supported but you would need a big fund to support a worthwhile amount to live on.

Yes, “pretty sure”. Exactly.   Assuming the next 50 years will be like the past 100 years.  

The word “safe” in  SWR leads people to be believe that its a deterministic guaranteed value.  And it isn’t. Its a probabilistic guesswork based on assumptions.   

JohnWinder

Yes, it's worth reading more fully on this, as one would discover that 'safe' means the method fails in only 5% of circumstances based on history. That wouldn't be safe enough for me, but you only have drop to about 3.7% instead of 4% withdrawal to be 99% certain of never running out. And if one kept reading one discovers it's not designed for 50 years of withdrawal, but 30.

Linton

JohnWinder said:

Yes, it's worth reading more fully on this, as one would discover that 'safe' means the method fails in only 5% of circumstances based on history. That wouldn't be safe enough for me, but you only have drop to about 3.7% instead of 4% withdrawal to be 99% certain of never running out. And if one kept reading one discovers it's not designed for 50 years of withdrawal, but 30.

Also the figures are based on the performance of US investments over the past 150 years and are believed to be lower for UK investors, particularly those who invest in the UK which as been the majority for most of that time.   On the other hand the past 150 years include 2 world wars, several technological revolutions, the rise and fall of empires and major economic powers etc etc which may or may not be relevent to the next 50 years.

So all in all I think SWR figures may be useful for a sanity check but cannot provide any sort of guaranteed basis for one's entire retirement.  Certainly anyone who says that their plan has a x% chance of future success is fooling themselves.

Terron

SWR are not based on "probablistic guesswork" but sound mathematics, historical data and the assumption that the future will not be worse than the worst of the past for which we have records. Given that the period covered includes the great depression and the oil crisis that worst was pretty bad so it seems a reasonably sound assumption. It also includes two world wars but they were good for US investors. The original work was done using US market data but has been repeated for UK data (giving a lower rate of ~3.5% IIRC). The original work assumed that a 5% risk of funds not lasting 30 years was safe enough, but has been repeated for longer periods and lower risks. Going for 0% risk would be stupid, but 1% is not unreasonable. Of course innumerate people may misunderstand the term SWR and there are far too many of them around, but that is not a problem with the term.

The most important thing about SWR IMO is its inverse. If you consider 4% to be safe enough and you know the income you want in retirement (your number) then you have a target for pension savings of 25 times your number. That is simple enough to use a rule of thumb to set an initial target for your pension fund.

michaels

Two little known issue with the SWR:

1) It effectively assumes perfect foresight during the drawdown period - in some scenarios about half way in the pot gets down to only 3 or 4 years worth of withdrawals but a market boom makes it all work out in the end.  In reality if you were living through that 'sequence' you would very definitely have cut expenditure before things got so tight hence actually you can't be certain that you would really take a constant  inflation adjusted' income despite what the history tells us.

2) Constant real terms income actually falls behind average earnings over a long retirement - you could quite easily retire on 'average income' and after 20-30 years be below the poverty line.