If an SWR is just that, how come time of retirement can make so much difference?

Audaxer

Deleted_User said:

IMHO SWR is mathematical nonsense because it assumes a constant withdrawal  from a highly variable pot. It’s counterproductive but heart-warming therefore popular. 

I agree there is no guarantee it will work in the future, but as I recall it shows that a starting withdrawal of 4% rising each year with inflation, has worked in around 90% or more occasions over a 30-year retirement. I therefore think it can be a useful guide for retirees looking to start drawdown.

GazzaBloom

I prefer to see SWR as "starting" withdrawal rate. I am considering 5% starting with a 90/10 or 80/20 stocks/cash mix.

We have to start drawdown somewhere and what will unfold after that is unknown.

michaels

GazzaBloom said:

I prefer to see SWR as "starting" withdrawal rate. I am considering 5% starting with a 90/10 or 80/20 stocks/cash mix.

We have to start drawdown somewhere and what will unfold after that is unknown.

If you go high equities even though the total return will be higher on average you are likely to see more volatility so more likely to have to make big cuts to your initial annual 'income'.  Presumably considerable annual volatility doesn't worry you.  The principle of the SWR is it is also suggested to be an amount you (based on historic performance) could take (with RPI increases) for ever.

There are lots of potential other strategies if you are happy with fluctuating income, but they don't do what SWR 'should' do.

OldScientist

DT2001 said:

OldScientist said:

NedS said:

Given many global equity funds are yielding around 2% (e.g, VEVE), and 4-4.5% is available from gilts, a balanced portfolio may give you much of the income you need from yield and you may only need to realise a small proportion of income from selling units. If 3% of your 3.5% SWR is coming from yield, you shouldn't need to sell much.

The problem with the natural yield approach is inflation - what might be 3.5% this year, will, in the absence of capital growth (which will be absent for bonds since the coupons are being spent), in real-terms be worth 3.15% next year (with current 10% inflation) and so on.

Abraham Okusanya backtested this approach with UK data (the results are at https://finalytiq.co.uk/natural-yield-totally-bonkers-retirement-income-strategy/). While I disagree with his use of the word 'bonkers' since the income from natural yield is no more variable than that for any other percentage of portfolio approach, it would require a retiree to have a significant amount of their essential spending covered by guaranteed income such as the state pension, DB pension and/or RPI-linked annuity or, as he says, to have a sufficiently large portfolio such that even in the lean years it still provides sufficient income to live off.

I have copied and pasted (below) a response to your linked article and having read some of your other posts see that you are well researched and wondered what your take was? My MIL has an income based portfolio (built to utilise the IHT relief of passing on excess income) and although it has only been in existence for 5/6 years it seems to have produced steady returns whilst the capital value fluctuates.

If one can combine some fixed income (maybe through fixed term annuities), DB and SP you can generate steady but unspectacular returns whilst protecting yourself against SORR?

This article is incorrect, both in theory and in explanation.
The practice is not about yield producing, but stable yield producing, which is where an intelligent adviser starts research. Running back tested portfolios from 1900 is a pointless exercise, and bears no relevance to anyone alive today, or their parents.

Investing for long term rising income is the hardest task for an investment manager to undertake, however we do have the luxury of a swathe of long term divi producers to help out. Since Procter & Gamble produced their first divi in (I think) 1897, they haven’t skipped one, and in the last 60 years it has risen every year. Closer to home, City investment trust was the first animal over the 50 year hurdle of consecutive rises, and this year should see Bankers and Alliance Trust join them, Caledonia next year. These are not one hit wonders, and they all have competent revenue reserves to back up their payments. 

the leading reliable investment trusts had an average payout increase over the 15 years from Jan 2000 of over 6% per year – that’s the annual increase, not the yield.

Yield to cost.
An investor receives an income that is his/her yield to his/her cost. Yield to current share price is only relevant to today’s purchaser. (Read anything Buffet, Malkiel or Ellis have ever written on divi investing).

Example:
Murray International. A purchaser in Jan 2000 bought a share with an existing 3.18% yield. An investor buying in 2014 was buying into a 4.38% yield. However the 2000 investor’s 2014 yield was 9%. The share price paid in Jan 2000 was £5, the divi received in 2014 was 45p. The focus on current yield is maintained by those in the markets trying to encourage trading, and the luxury that an adviser has is to evaluate the investment, and income, for the individual client and their own, personal, unique portfolio.

City of London’s divi in 2000 was 7.18p, and in 2016 was 15.9p. Share prices were £2.42 and £3.79 respectively, meaning the 2000 investor last year had a yield to cost of 6.57% – and capital growth of 56% which we are ignoring.

It’s important not to take investment theory designed for institutional mandates and apply them to retail clients – the two cashflows are entirely different, not least because on retail we have the luxury of finite drawdown periods, and a known investment timescale. David Swensen, in his time at Yale, did point out that investing in ‘alternatives’ was very relevant to long term investing for them, in fact a key ingredient, only because the endowment’s time horizon was 150 years.

“a natural yield portfolio will specifically overweight high-yield assets…”
Not true. High yield necessarily means ‘higher than the mean’, which to us simply means pushing the edge of the envelope, and that is anathema to long term income investing. The focus is not on high yield it is on consistent yield. Proven sustainability is key, and that can be researched and assessed via the published accounts, this year, last year, the year before….

When anyone in our back office team starts to get bogged down in the ‘science’ of investing I tape a pencil to the top of their computer screen – better than a pen in space.

The income investment equivalent of the Russian pencil is the London listed investment trust. If you can buy City of London today on a 4.2% yield, and with annual growth of over 5% over the last 16 years, which part of the objective is not being met?

More importantly, which of the two methods do you fully understand, and which of the two will the client understand?

I think that looking at the performance of different approaches using historical data is relevant since the testing then covers a wide range of conditions - some of which may never happen again, but some of which can resemble current and future conditions. For example, retirees beginning in 1900 had to cope with high inflation (caused by WWI), a pandemic, and then deflation (we've recently seen the first two of these, but not the third). However, using the past 15-20 years as an exemplar runs the risk of recency bias.

If I've done my sums right, it is interesting to note that for retirements between 2000 and 2016, the income derived from the dividends from a portfolio solely consisting of the FTSE100 would, for a £100k portfolio, have ranged from £2000 to £2800 in real terms (i.e., referenced to the prices in 2000) and the portfolio would have been worth about £75k in real terms at the end of that period. That is quite a poor performance compared to the quoted Investment Trusts over the same performance (albeit neither holds entirely UK shares - City of London is currently about 80% UK, while Murray only holds about 8% UK), so the comparison is not entirely fair. My (limited) understanding, is that the gearing available to investment trusts can be used to help smooth dividends in turbulent times and to provide additional growth.

Without going down the investment trust route, if you want to to smooth potential variations in a percentage of portfolio approach (whether that is natural yield, constant percentages, variable percentage withdrawals, VPW) your suggestion of building up guaranteed income is a good one (with single life RPI annuities now at over 4% at 65 years old, they are certainly more attractive than any time in the last decade or more). With VPW, the Bogleheads appear to have settled on using a cash buffer to smooth the withdrawals.

Finally, for disclosure, I am firmly in the total return camp (I'm applying a version of VPW to my own portfolio withdrawals, but I also have a DB pension to cushion variability), but I can see the attraction of the natural yield approach for its simplicity and, if you pay for transactions, a reduction in costs. In the past, when stockbroker fees were significantly higher, this was of particular importance (e.g. my father's retirement portfolio was completely set up to deliver regular dividends and coupons - the latter with an attractive 8%).

GazzaBloom

michaels said:

GazzaBloom said:

I prefer to see SWR as "starting" withdrawal rate. I am considering 5% starting with a 90/10 or 80/20 stocks/cash mix.

We have to start drawdown somewhere and what will unfold after that is unknown.

If you go high equities even though the total return will be higher on average you are likely to see more volatility so more likely to have to make big cuts to your initial annual 'income'.  Presumably considerable annual volatility doesn't worry you.  The principle of the SWR is it is also suggested to be an amount you (based on historic performance) could take (with RPI increases) for ever.

There are lots of potential other strategies if you are happy with fluctuating income, but they don't do what SWR 'should' do.

I should have added that it's only around 5% for the first 9 years then drops significantly when SPs kick in, my biggest challenge is SORR or sudden death/ill health. 

Terron

Deleted_User said:

IMHO SWR is mathematical nonsense because it assumes a constant withdrawal  from a highly variable pot. It’s counterproductive but heart-warming therefore popular. 

No, it is not nonsense. It is an attempt to emulate the gold standard of pensions - a fully indexed DB pension - using a DC pot. There may be better alternatives, but that was a sensible place to start.

Cus

Terron said:

Deleted_User said:

IMHO SWR is mathematical nonsense because it assumes a constant withdrawal  from a highly variable pot. It’s counterproductive but heart-warming therefore popular. 

No, it is not nonsense. It is an attempt to emulate the gold standard of pensions - a fully indexed DB pension - using a DC pot. There may be better alternatives, but that was a sensible place to start.

But DB pensions are invested in a similar way to annuities, in that the providing firm has to ensure you are going to be paid for as long as you need. So for me, the only safe withdrawal rate is likely to be similar to what you can get from the annuity market.

It is heart warming to think you can take x% and most scenarios play out in your favour, but if it was that simple, firms would offer annuities based on equity investments, but they can't/don't.

bostonerimus

SWR depends on a large number of parameters and crucially many assumptions. In fact SWR is a large set of numbers depending on those parameters and assumptions. So if you use historical data and historical average inflation you'll end up with a set of SWRs depending on the sequence of market returns and annual inflation numbers. There will be other numbers at either end of the spectrum that use either optimistic or pessimistic numbers. And then there's the critical factor of longevity which again has a spread. So you have to approach SWR really as a set of SWRs with associated probabilities. A common approach is to choose only the SWRs that have a 95% probability of success ie not running out of money.

MK62

There is no such thing as a safe withdrawal rate for future retirees.........all there is, is a withdrawal rate which would have been safe for past retirees....given the individual retirees set of circumstances.
That said, you have to have a starting point, and what other references are there?.
As several have said though, it should be seen as a starting withdrawal rate, rather than safe.......it could be that a new retiree will need to change their withdrawal rate in the future, but as long as that's understood at the outset, what better way is there for a new retiree to decide on a reasonable withdrawal rate?
If you want more "safety" than that, then an annuity is probably the only option that will deliver that goal, but the downsides to that are that the rate might be lower then the SWR above and your "pot" is gone........in the end though, it's a judgement call each retiree with a DC pot is going to have to make.

Audaxer

I think most retirees that start on a 4% withdrawal rate will use it as a guide, and most will have a cash buffer to stop them drawing down from their portfolios during equity crashes. I have read than in most cases retirees end up with a bigger pot than they started with, due to them taking too safe an approach.