Dynamic asset allocation and withdrawal in retirement

Many thanks for this – a really useful post.

If I could ask a question – why don’t you include the State Pension as part of your Plan A – rather than as a backup plan? Too much political risk?

@Dawn:
Re: “I’m 55 ,my equity, cash, …..”
If I have understood you correctly, this is an interesting scenario. Also, I assume (in due course) you will draw a state pension too?
Does this mean you may still actually be accumulating, ie your net assets are still growing?
If this is the case, may I ask how do you handle any surplus?
I ask because all decumulation models don’t really cover this scenario.

Another great one in the series. Many thanks.
The length of time before the state pension kicks in and gives you a handsome pay rise seems to be the big reason for your conservatism and a plethora of back up plans perhaps? I like that as it has the same belt and braces approach that I take.
Our time at planned retirement (total retirement if we want it) to state pension hike is half(partner) at year 5 – and half(me) at year 9, I view these as milestones and have designed our draw down of assets to cover the shortfall from our two DB schemes accordingly. This, by it’s nature, has a slightly high SWR for a few years, but a lower one if necessary later after yr5 and 9. Worst possible scenario is work or downsizing/equity release, which we are reluctant to do as we quite like the house and so does our singular offspring who has her eye on it 🙂
The idea of a dynamic approach to managing the assets in draw down will give me a little further reading to do I have read some but your references tell me I should read further. Idealy I want to be done with thinking too hard about money by the time I’m 70.
JimJim

Happy to see Part 2 of your @TA’s master decumulation plan. McClung is indeed the Man ! and the backtesting looks compelling. Personally being very much a buy and hold investor, and migrating years ago to having around 80% of portfolio in Vanguard LS funds, i dont feel like switching to separate Bond and Equity holdings to implement McClungs dynamic asset allocation model is for me. Rather i will use the 3 years baseline cash that i hold for living expenses, and top it up by selling equities when annual portfolio returns are circa 5% or more up. This may well not squeeze out as much return as the dynamic allocation method but its simpler to operate, am i missing anything ??
Re Dynamic Withdrawals: Agree totally on being able to flex spending down from baseline SWR and have significant room to do that. My plan is to look at all 3 measures (base SWR (3.5%) + inflation, McClungs xls, bogle heads VPW xls) each year to come up with a “budget” amount for the following year which will be somewhere between the 3 measures based on how we feel about mr Market and what plans we have for spending/travel in the following 12 months – its a kind of sitting on the fence approach cos for a 30+ year retirement i dont feel i want to commit religiously to any one single method – Again thoughts what am i missing ??

again really looking forward to following the debate and comments on this post its supper helpful thanks @TA and everyone else for sharing !

Interesting. Let’s say (for whatever reason) you decided to just use Lifestrategy 60 for your drawdown. Would you aim to take 4% a year withdrawal – lump sum? (keeping 2 or 3 years as a cash reserve – as a backup). Does anyone do this? Simplicity seems very attractive at times. Oh what to do…

Loving this mini-series. It’s just interesting to compare what we’ve been doing with how you are thinking about it all.

I have to admit, I tend to think ‘systems’ are over-rated to some extent. To a large degree, it’s just common sense? Don’t be forced into selling something when you don’t want to. Create enough diversity & flexibility in your portfolio to adapt as needed.

Same way on the budget front, as you mention. We ended up with way more flexibility on that front than expected, largely due to our travel plans being much cheaper than we thought. Slow travel is very different to holidays!

So yep, you need a plan/approach – but the biggest thing by far is being able to adapt as needed. And have plenty of back up plans – which I look forwards to reading about in part 3 🙂

@A1Cam

I hit my FI number 2 years ago so decided to drop to partime work. Yes, and let my untouched portfolio keep growing . I dont have the same surplus cash available to invest now on a partime income but I’m finding I’ve still got money over spare at my year end.! Once a saver, always a saver. So i tend to reinvest it in SS isa, cash or premium bonds. Last year i spent some on a small cosmetic surgery job. But if I’ve none spare at end of year its OK. I’ve done my grafting and saving. I did think about spending going on more holidays too as I’ve got more time now and I’m fit and healthy.. Yes, i expect to get full state pension in 12 / 13 years time which will be a fantastic boost @8k but I’m a tad nervous it may be cut or means tested so carnt rely 100% on that boat coming in, that’s why I decided my SWR would be 3.5%. On my portfolio. I’m currently getting a 4.7% return on my rental property. After I’ve put aside repairs maintenance money.

@TA – Excellent article as always, thanks!

I’m also considering going for McClung’s prime harvesting but I am tempted to choose a more conservative safe withdrawal rate at least in the first couple of years: (1/PE) * percentage of stocks + 1% * percentage of safe assets – 2%. This 2% is a rough estimate of the average inflation rate l would expect in the near future. My portfolio has a net Equities PE between 16 and 17. I end up with a safe withdrawal rate of about 2%. If all goes well for the first couple of years (or for the first 5 years perhaps), my plan is to increase this SWR to about 3% and use McClung’s Extended Mortality dynamic withdrawal method.

@ ZXSpectrum – “Millenials and Gen Z will outnumber Boomers and Gen X by the end of this decade.”
Couldn’t agree more with your point, and this is also why I think the state pension should not be relied on too much even if you are only about a decade away from receiving it.

@ZXSpectrum48K (17) I do hope you are wrong, (And you are right about so much) – But, our demographic is still top fat with life expediencies static or improving depending on who you read (Improving especially for the wealthier end of society) – also remembering that in ten years time millennials will be thinking about their own pension schemes and voting to cut that effects them too – also people vote more the more they feel enfranchised by the state, again, the wealthier end of society… Could it happen – yes, undoubtedly – but as we do not stack up well against many other developed countries for state benefits in retirement and the political will has been to rectify this for some time now, I think the risk is small… Not that I don’t have a plan to mitigate that small risk 🙂 As for DB schemes, they have been eroded for many years now in the public sector and I can see this being the method of reduction into the future rather than going back on a promise to pay or abandoning them altogether. DB schemes in the private sector are now like gold dust.
JimJim

@MrOptimistic i am quite aware that blended LS funds dont suit McClung or other decumulation strategies that require separate Bonds/Equities to operate, they do however offer simplicity which may be something myself and others value, admittedly at the cost of future returns if McClung is correct

@Dawn:
Thanks for your detailed response.
I suspect that this ‘problem’ is possibly more wide spread than people may at first assume.
I summarised my situation in a comment to the first Part of this series as “I am some four years plus into ……….. and have found that I have been quite reluctant to push any surplus (vs planned spend) directly back into equities.
What I have done instead is up the equities exposure of my upside Pot. However, this can only continue for so long.”

I am currently not sure which is the best way to proceed in the mid/longer term, but it surely must rank as one of the better problems to have.

I have yet to read the free chapters from McClung (note to self (again!) as I downloaded some time ago) but after reading your 4 bullet-point precis I have an itch that I am hoping you or one of the readers can scratch. (Hopefully it is covered in the book.)

Bonds are sold for annual expenses, rebalancing is from equities to bonds only provided there has been sufficient growth, and if bonds run out then equities are sold to fund annual expenses.

My itch is a scenario where equity growth is weaker than the rebalancing threshold, all the bonds are sold off from year to year, and then horror of horrors, equities crash just when you start to sell them. Perhaps I am overly pessimistic but does McClung consider this possibility? And what is his ‘Plan B’?

I am aware that you have your ‘plans in depth’ (https://monevator.com/how-to-protect-your-portfolio-in-a-crisis/) and I am looking forward to seeing Part 3 and your plans.

@ZX…interesting point ab out the direction of population dynamics…how might HMG have a go at rebalancing against the DB brigade ? As many more in a few years hence will no longer be contributing to them , ie deferred or in receipt.

It is interesting to see the concern about the future of the state pension here as, in the comments to part 1 in this series, everyone (IIRC) appeared to be in favour of buying extra years and/or ensuring maximum state pension payout but there seems to be less positivity on these comments.

@MarkR #26
The whole point of a withdrawal strategy is that plan A should work well, and you don’t need to change strategy half-way through? The scenario you describe is difficult for any strategy, and with most strategies you would have had to sell stocks even earlier. Or do you have any alternative in mind?

McClung himself says for example: “Prime Harvesting can sometimes have a seemingly undesirable side effect: during long periods of low stock returns, bond levels can go below their preferred limits […]. However, Prime Harvesting has a strong counterbalance to these occasional low bond levels – these periods correspond to attractive stock valuations. […] [A]ttractive stock valuations substantially reduce exposure to known risk. It’s important to understand that growing stock percentages using Prime Harvesting is not an anomaly, but a response to the market, trying to minimize overall risk.”

@Mark C
I’m more sanguine about the State Pension continuing to exist than some some others in this thread. But nothing in this life being certain, topping up your contributions might best be seen as just another way of diversifying your portfolio and spreading risk.

Living the Deaccumulation!
Retired at 57 -17 years ago-now nearly 75
Got McClung plus as many others as I could read-still reading but it’s now getting late and course set seems to have worked
Used Bogleheads Dynamic withdrawal method-very simple
I enjoy finance but I do have other things to do!
Withdrew much the same amounts that gave me equivalence to to my working salary
This turned out to be right as wife and I travelled a lot -more free time
A set Asset Allocation 40/60/5 equities bonds cash- 3-3.5% Withdrawal rate
Quicken 2004 has small Portfolio rebalancer program which tells you how much required from equities and bonds to retain Asset Allocation integrity
One withdrawal per annum to top up the 2 years expenses cash float
Now at 30/65/5 -time to start increasing equities again ?-if I can raise the enthusiasm
3 index tracker funds only
Simple cheap and easy to follow
So far so good!
xxd09

@TI – a retiree wouldn’t be a non earner because of sipp and state pension income, so might have access to conventional finance like 0% cards, interest only mortgage, maybe offset mortgage, even equity release, etc, and they might carry forward these facilities from pre retirement. I wouldn’t touch margin loans and I regard standard unsecured loans as mostly pointless with a 5 year term.

Just mean that selling down other people’s bonds has similarities to issuing your own. 0% cards are generally good because you can (usually) keep kicking the can down the road although they won’t cover your expenses for long

There could indeed be times when taking on debt to preserve equities was bad but assuming a similar interest rate I imagine they’d be the same times when dynamic drawdown wouldn’t work (ie bonds outperforming equities)outperform

But yes, peace of mind important. In reality I dont need much and could afford to live off the state if I had to. I have trouble finding things that I actually want to buy (happy with old banger, prefer hone over holidays, dont apprentice posh nosh, etc)

@TP2, have you set aside a “care home fund” from the beginning? Our thought was that our initial extra expenditure on travel etc (when permitted) would start to reduce and allow us to increase the emergency cash fund for that purpose. Or just stay in investments that are reasonably liquid.

And those extra costs you mention do still exist. Yes we paid off the morgage when I retired, but we have a daughter at university who needs parental contributions. And no doubt at some point, unless property prices change dramatically, she is likely to require parental help if she is to get on the housing ladder.

@JDW:
This is not a direct answer to your NI question but might help with your thinking.
I am still under 60 and pulled the plug a handful of years ago. Due to being contracted out, I am currently short of the ‘years’ needed for full SP. I need to fund a few more ‘years’ before I turn 67 to get full SP.
N.B. part ‘years’ are not a purchasing option although they are a pay-out possibility, see more at point d) below.

My current plan is to wait until nearer 67 before parting with these funds, because:
a) I paid for an extra ‘year’ under the old scheme – which then counted for precisely zilch when the system/rules changed,
b) I might (stress might!!) do some paid work between now and turning 67,
c) there is no longer any survivor benefit under the new scheme, and
d) I may not buy my final missing ‘year’ as whilst it costs the full amount it will only provide me with part of a ‘years’ state pension pay-out

However, things do change so I keep the situation under review and am fully prepared to change course if necessary.

IMO such a degree of flexibility is probably necessary to successfully navigate de-accumulation too. That is, whatever system you currently plan to follow, at best, it can only provide near-term guidance and must regularly be reviewed and revised as events unfold.

@Al cam, re your final year missing year, Ive just discovered Im in a similar boat – Im short one year, but at a (current) cost of £795 I believe, I would only get an additional £28.40, which is what I’m short of the full NSP. I assume this irregularity/small shortfall cane about because of the calculations to take account of being contracted-out some years ago. Glad I checked with the pension service.

I think this series will join the ‘how much do I need to put in my SIPP and ISA’ from early 2020 as a must read. Thank you!

Like @Al cam, I’m interested in how to combine this with an ISA bridge. I’m likely to have a 12 year bridge and was thinking of a 40-60 stock bond allocation, rebalanced annually. I read Wade Pfau’s ‘How much can I spend in retirement’ which has a handy table showing the SWRs for different asset allocations and time periods (I think it was 8.2% for 12 years). So, I was just going to use that

I suppose there is no reason why you couldn’t use prime harvesting for the bridge and then the dynamic asset allocations and withdrawals once you can access pensions.

@JP:
Just to be clear did you mean £28.40 per annum or per week or what precisely?

@TA
Another great article. I bought McClung’s book last year after reading your review, it’s very comprehensive and insightful. I like the Prime Harvesting concept and McClung does a good job of demolishing some of the alternatives.

I get the point made in your comment that there are multiple redundancies built in to your approach, and I expect you got comfortable on that basis. But like a couple of other commenters I was wondering if you had considered a more direct way of attacking the sequence risk problem. I’ve spent a bit of time looking into trend following (as mentioned by Mike at comment 5), which does seem like an interesting way of potentially reducing volatility without harming returns too much. There is a little favourable research on this (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2764933) but generally it doesn’t seem to have received much attention in the retirement context.

Of course it seems like anathema to us buy ‘n hold types, and as it would need (at least) monthly checking of asset values against 200 day SMAs or whatever, perhaps it’s just too much trouble for a retiree – even an early retiree.

There are even funkier possibilities, like tail risk hedging, that I haven’t looked into in any detail. I do think that there must be more focused ways of dealing with the sequence risk problem than just relying on a stocks and bonds portfolio, even if tied to a variable withdrawal/asset allocation strategy. Perhaps in time the investment & pensions industry will make some sensible options available to the consumer, though their record on retail structured products is… not good.

Great series so far and already looking forward to the final episode…

I’m 5 years into retirement now, though nearer a traditional retiree (I was 60) than a RE type. Like TA I’m most keen on the glidepath decumulation strategy outlined in Michael McClung’s book, but moderated by EarlyRetirementNow’s critique which identified a couple of very odd boundary behaviours of the raw prime harvesting method. In my reading of ERN’s modelling it seems a consistently increasing glidepath up to 100% equities is as effective or better than PH. I get lost on how this glidepath idea applies to retirees though, as ERN is dealing with the run up to and through retirement, despite going head to head with McClung. ERN says post-retirement things are different and you need chunk of cash and/or bonds for long term security and the ability to sleep at night as much as anything.

Now in practice what has happened is I’ve followed the McClung PH method alongside ERNs models (he also has a free spreadsheet you can tinker with) AND an annual actuarial-based view from howmuchcaniaffordtospendinretirement.blogspot.com/ – they provide an ABC spreadsheet. I changed the level of equities I was happy with (from about 63% to currently 73%) so that has thrown the PH spreadsheet up in the air – I’ll have to figure out how to restart it or otherwise muddle through. I changed from an annual drawdown into a high interest easy access account to a monthly one (that’s ERN for you), but draw down the CAPE-based estimate (but ‘mentally’ account for the extra I could have drawn down by adding the difference with that to the plain version (about 10% more) and ‘save;’ it in a daft expenses account for holidays and such like, so I will feel less bad if I do some of that). I keep an eye on the ABC actuarial valuations (usually MORE generous than the ERN CAPE based ones, but my assumptions are maybe different between the cases).

And so on. It’s human instinct to try to divine the tea-leaves and agonise about stuff that is inscrutable, and sadly I’m only human.

My real safety cushion is I have a good DB pension and have just paid off a few extra years of NI – my state pension kicks in in October. Three-quarters of my income needs are going to come out of these resources. On reflection if I had to fund the lot out of a portfolio in freetrade I’d probably be somewhat less sanguine about how much to put in equities – really TBH I can’t see why I don’t just go 100% equities. You don’t often see commentary about how to modify drawdown scenarios for pensions and annuities – ERN’s spreadsheet does this though as does the ABC spreadsheet. I think perhaps if I got to 60 and had virtually no pension but just a big sack of cash, say from selling a business, I’d be tempted to cover my core expenses with an annuity, despite the epochally low rates, just for peace of mind, simplicity and security against longevity risk. I think this is the kind of approach Wade Pfau might advocate. And with the Grim Reaper and all that round the corner one has to face up to simplification and the security of ones finances earlier rather than later. https://monevator.com/death-infirmity-investing/

Probably past time we heard from the Greybeard again. Hope he’s OK.

@Al Cam, yes, should have made that clearer – £28.40 per year. A very small shortfall.

@ Al Cam & Ray – I will be using a cash bridge for over 6 years but that doesn’t exclude running the rest of the decumulation plan. I think of cash as part of my asset allocation, not a beast apart. McClung’s method calls for me to use cash / bonds for living expenses. Equities rebalance into bonds after breaching the 20% inflation-adjusted increase. I’ll follow that system once decumulation starts.

@ Dawn – it sounds like you’re nicely set up. As you illustrate, every solution comes with trade-offs.

@ Mike & c-strong – I have similar thoughts to you about trend following. Plenty of evidence in its favour but it does require you to stick with it through hell and high water. And there seems to be plenty of high water. You can lag the market for years before making it big. Given my own psychology, I decided to get exposure to momentum via the multi-factor allocation where it has a low to negative correlation with value.

@ SirRik – my mention of Vanguard LifeStrategy in connection with cognitive decline also means buying an annuity. The annuity would be the ‘bond’ side of the allocation in that scenario.

My assumption is that this could happen when I’m much older. Annuitise as much of the portfolio as necessary to provide a solid floor in an index-linked product. The remainder of the portfolio goes into Vanguard LifeStrategy 100 – but only if that money isn’t really needed. You live on the annuity while the equity side of the portfolio is free to grow and is essentially on a rising glidepath.

It’s a very simple approach that I’ve seen work wonders for a close relative. The annuity takes away day-to-day stress but equity upside is still a possibility. They’re not interested in anything more complicated than that.

@ Cigano99 and Hal – holding a set number of years of cash is a bucket strategy. There’s been lots of research into it, and that typically shows that a baseline level of cash proves to be a drag over the years and doesn’t offer any more safety. That said, I think it’s far more important to be comfortable with your system than optimising everything to death and experiencing on-going, low-level stress about it.

@ Doris – what’s the alternative to one day selling your equities? I think it’s important not to anchor on any given equity valuation. As long as the portfolio funds a long and happy life then that will do.

@ John Tickner – I signed up for the free trial here:
https://app.timelineapp.co/registrations/new

@ Tom-Baker – The lower your SWR the safer you’ll be. I’m too impatient to wait any longer

@ Ash – that reference is meant to highlight that mortality-sensitive systems recognise that you aren’t going to live forever. In other words, if you knew you were going to live one more year then your annual withdrawal could be 100%. Vanilla SWR rules only allow for a standard inflation-adjusted increase even if you’re 110. They don’t take into account your attenuating lifeline.

That said, I don’t buy into this idea of declining spend as you age. Even if the research is right on average I’m not banking on it applying to me. That’s just another of my conservative assumptions.

As an aside, we talk a lot about care homes, cognitive decline, failing health etc on these threads. I intend to kick against this outcome as much as I possibly can.

@ MarkR – Choosing a conservative SWR is the central brace against such a fate. You’re setting your withdrawal rate low enough to not exhaust your portfolio based on how those scenarios have played out in the past. McClung models a number of Plan Bs such as back-up portfolios.

@ Whettam – I did consider a rising equity glidepath or bond tent. I like it. Ultimately I decided that the dynamic withdrawal rate is more flexible, and achieves much the same objective re: safety. Only history will tell whether I made the right choice.

I’ll maintain the asset allocation outlined last episode between bonds, cash and gold. The caveat is that gold gets spent down and never replaced in a crisis when bonds and equities are tumbling off the Reichenbach Falls together.

Re: Pfau and PMT – yes, I’ve read about it. Pfau is a great academic and I’ve found his work invaluable, especially on international SWRs and annuities. But in my experience, he’s never really offered a fully-featured, practical solution for DIY decumulators who can’t afford safety first. Are you considering PMT for your system?

@ Nearlyrich – cheers!

@ GMO – that’s a great review. The ’squeezing the balloon’ metaphor is a great way to visualise the problem.

@ IanH – thank you for your insight. It sounds like you really enjoy the process of being hands on, and the cushion is an important part of enabling you to experiment.

My reading of the research on rising glidepaths is that they ultimately perform best in a narrower range of market conditions than dynamic withdrawal rates. Much depends on your inputs and ultimate objectives though of course.

If you had to choose a system which one would you go for, or would you still create your own hybrid?

@TA:
And it turns out that in the real world Ian H (#54) discovered that feature of McClung (#59) and decided – IMO eminently sensibly – to be pragmatic!

@ Al Cam – I will be decumulating. But, this doesn’t seem like a problem to me either way. The decumulation strategy allows for the fact that the portfolio can increase in value. If that happened to be with new money then I could just buy the same asset allocation in proportion.

Or I could use it as a back-up portfolio – some people call this a fuse portfolio. McClung describes different variants. Or I could invest in a ladder of index-linked bonds, or annuities.

The SSiRS is new to me. Thank you for the link. Do you intend to use it? How do you rate it?

@ Al Cam – You’ve put your finger on an important sub-text of every comment thread on this topic.

If the ‘answer’ is to bombproof then there is no answer because there is no bombproof. We can up the Armageddon ante all day long.

If the ‘answer’ is a system that you’re comfortable with, understand, can operate and considers the variables you believe to be relevant to your situation then there is an answer. But it probably isn’t one single system. It’s a blend of considerations bent around your personal circumstances.

All systems have their shortcomings. It’s important to know what they are. It’s also useful to know where people are coming from.

If someone is getting 80% of their income from a DB pension and doesn’t care about leaving a legacy then their system may not contain many answers for someone relying on a defined contribution portfolio and for whom leaving something for the kids is mission critical.

Similarly, I can totally buy into the ‘keeping it simple’ argument.

I don’t think you can take a single factor like an SWR and say, ‘that’s it folks, nothing else to think about.’ Although it’s simplicity that’s the allure of the 4% rule and why it’s famous.

Can you outline what your solution will be? You’ve obviously thought deeply on the topic.

@TA
I suspect that my thinking on this is too simplistic, but isn’t it true that in the limiting case where equity prices fall monotonically from your day of retiral until your demise then the earlier those equities are sold the better? From my understanding, the proposed strategy would lead to the tail end of the portfolio being 100% devalued equities. I appreciate that no strategy is risk free, but if you’re in the camp that sees global equities as being currently very overvalued, might this not be a bad time to set the ‘base value’ for equities below which they won’t be sold?

@ TA Thank you again, for all the answers. I agree that Pfau does place a large emphasis on safety first for the income floor and this is not affordable for most and / or younger retirees. I have not decided my own system yet, which is one of the reasons I’m finding this series so interesting.
I ‘m attracted to the simplicity of the Excel PMT calculation for variable withdrawals. I think it has a number of advantages, including its simplicity, my wife has very little interest in all this stuff, I also like the way it includes current portfolio value explicitly in the calculation and incorporates changing estimates of mortality and expected returns. Ultimately I think I’m more likely to initially choose something like EM, but I will still calculate PMT as a “shadow” benchmark each year.
Thank you (again) @A1Cam for the SSiRS paper, I had not seen that either something else to read, along with re-reading McClung.
Although I’m totally convinced by Dynamic Asset Allocation, I do struggle a bit with the Prime Harvesting idea that you never buy more equities. Which is why I asked @TA about your thoughts on equity glide paths. Maybe it’s a behavioural / confirmation bias, I need to get over, but I have spent 20 odd years interested in Accumulation investing thinking about asset allocation / re-balancing. I think I’m likely to try and combine the equity glide path, partly based on a timeframe, but also some market valuation using CAPE. Then when I have reached max equities, then switch to McClung.
I know that this will be too active for some, but although my portfolio is operated on a total return basis. My system is likely to differentiate between selling equities for income and taking the dividend from equities as income and reinvesting dividends.

@Cruncher Hi I’m a similar age and similar issue. I ‘m now certain to exceed the LTA, although I plan to only crystallise up to the LTA initially, to delay the charge.
You should verify this (maybe speak with pension wise?). However my understanding, is you can crystallise part of your pension, as soon as you are 55, put the pension into flexible access drawdown and take tax free cash. When you do this, you can still pay into the part you have not crystallised, I don’t think you can pay into a pension in FAD so you would need to keep at least two accounts (which may incur more fees). My understanding is that, you don’t trigger the Money Purchase Annual Allowance £4k limit, until you actually withdraw funds out of your FAD account.
For all funds in FAD, there is then a second test against the LTA at age 75. I think this second LTA test is maybe a target for future chancellors, maybe to limit the value of SIPPs as an IHT dodge?

@TA, are you following the full McClung Prime Harvesting, or a modified version?

There were 2 aspects of PH I did not like and have changed. The first is that he reinvests all dividends. The only rationale I can find in his book for doing that is that it made modelling easier. I prefer to add them to the cash pot for spending.

The second thing I did not like was on selling 20% when the value of the equity investments rose by 20% real. 20% is a large amount to suddenly dispose of and my preference is to sell the excess over 20% instead of knocking the whole thing back down again. ERN included this “slice off the excess” option in his review of Prime Harvesting and found it performed slightly better than McClung’s PH.

@ Whettam – I think the idea that you never buy equities under Prime Harvesting is a misnomer. You sell bonds to fund annual withdrawals. Over time, you could end up 100% in equities. This is most likely to happen when equity valuations are low. The system is geared to preserve your equities from sale and allow them time to recover. As a proportion of your wealth, your equity allocation increases in similar circumstances as if you rebalanced into them. It’s just a different mechanism that’s meant to help you avoid catching a falling knife.

@ Naeclue – it’s hard to argue with ERN. But ERN wrote one article on Prime Harvesting. McClung wrote a book. And tested it against non-US historical returns. Therefore I’m likely to stick to the original.

Another point that becomes obvious once reading McClung’s methods comparison is that fiddling with any of the methods/parameters – even a little – will not give exactly the same results in backtesting i.e. it may behave differently for some sequences and starting conditions + sometimes more than you would expect. The stress testing chapters show this well where methods/parameters are pressurised and their “linear range” breaksdown somewhat. What looked good – ends up less attractive in more stressed markets.

*If* a parameter change increases sensitivity to particular sequencing then the results may be *materially* worse than expected (if that sequencing turns up again). Whether the magnitude of the change actually matters is a whole other discussion. The examples in his stress test chapters (on prolonged low variable income / failure definitions arguably do qualify as mattering.

The implication is that once you fiddle with the parameters then you really should rerun your own tests. You are no longer riding on McClung or ERN’s backtesting coat tails. As is his wont and indeed right ERN tests what *he* likes (and not necessarily an author’s exact intent or original setup). Where he errs occasionally is in characterising his test as a valid debunking of something he has himself adjusted. He did the same with GuytonKlinger. You don’t reproduce or invalidate anything by testing something which is *similar* but not identical. You may think your fiddling is immaterial or an improvement and it may or may not be true. ERN is great in the sheer volume of drawdown topics and myth busting he has covered. His drawdown series is a great resource.

Like some other commenters – I also find the prime harvesting tuned asset class sales/rebalancing thresholds counterintuitive and arbitrary seeming. McClung’s great help is his testing of what he proposes with more than one geographic data set so you are not seeing a tuned for historic US only story. For me the out of band testing is the extra contribution he makes over most other books.

This counterintuitive magic formula numbers point comes up a lot in machine learning. When you train a mathematical approximation for complex inputs to generate a predictive output against a data set – the somewhat predictive magic parameters appear in the trained “formula” which are not readily understandable to the lay person. This traceability of decision making being one of the remaining hard issues with the socially acceptable application of datascience/ML/AI more generally.

@TA
One last attempt to refine my thoughts, then I’ll leave it to those who actually know what they are talking about.

Your fourth rule is
‘Rebalance into bonds only after a significant run-up in the value of your equities.’

I suppose my question boils down to run up from what? Is it the equity valuation on the day you retire or from some temporal average of their value leading up to retirement? If it’s the former, is their any sensitivity over the next say 25 years to whether someone retired pre covid in February 2020 or a month later in the depths of the March 2020 trough? It feels like that could affect how many equities are being held further down the line.

@Ian H:
Zwecher’s book is written for the US market. Having said that, it is not too tricky to do read-across to the UK where applicable. IMO, the book, although at times not the easiest of reads, is far and away the most comprehensive exploration of Floor and Upside.

It’s interesting that although lots of researchers have covered the ‘dividend fallacy’, ‘why dividends are a bonkers retirement strategy”, etc. I don’t believe any have looked at total return strategy, but treating the dividends differently to the capital, when looking at withdrawals.

Very simplistically if a retiree has a horrid sequence of return early in retirement and spends their bonds first, these will last for X years. But if their equities are a producing 1-2% dividend and this was withdrawn as income, rather than re-invested, it would make the bonds last for longer. I’m surprised no-one seems to have investigated the value of the extra time vs. loosing out on the re-invested dividends.

@gmo I’m a data scientist (I confess I mainly manage these days) and I agree with your comments about traceability, but for me traceability of data and the ultimate decision is a must have requirement, for anyone in my team implementing ML Models within decisions. I think the harder issue is explainability of ML models to lay people and that’s why a lot of production models are based on well understood techniques like regression / credit scorecards. A lot of data scientists are not very good at explaining more complex methods and lay people are usually not great at stats! There are lots of new methods that can assist with this, but some people still argue about the Monty Hall problem or the Birthday paradox, because we humans are full of cognitive biases 🙂

However I don’t think this is the issue with retirement research, we are not talking complex models, as @TA says “given the sensitivity of retirement research to inputs”. This for me is the real issue if you use historic data and want a 45 year planning window, then you do not have many time periods to look at and the world has fundamentally changed since the data used in some of the research. Three points don’t make a trend.

@TA – I agree that seeking precision is misguided. We can’t predict the future, merely horizon scan for potential problems and try to mitigate any adverse outcomes. “Am I about right?” is as good as it can get.

Personally, I think we might be about to enter the Great Stagnation. Asset prices have been boosted to the max (it’s like a drug Govts can’t kick – “got some nice Stamp Duty holiday here. Very pure.”) and Governments have very little ammo left – maxed out on borrowing and interest rates lowest in 300 years where is there to go? Added to that, increased protectionism seems everywhere a risk, inflation may pick up and unemployment has been delayed by massive Government borrowing.

But I know nothing. Ho-hum.

@Brod, etc

To quote Carveth Read, see e.g. https://en.wikipedia.org/wiki/Carveth_Read,
“It is better to be vaguely right than exactly wrong”
This quote, or some subtle variant thereof, is often attributed (incorrectly, I now believe – thanks to a previous chat on this subject at Monevator) to Keynes.

@Naeclue Your approach is interesting thank you. I agree it’s harder to research, but it seems to me such a fundamental factor with equities I’m surprised that someone has not tried.

@Brod I agree with principal that spending the dividend is like ‘selling’. But in practice dividends have not been as historically volatile as equity prices, so spending just the dividend and hanging on to the capital, will have a fundamentally different outcome from a deaccumulation performance perspective. If we take a fairly standard Dev World ETF like HMWO:

https://www2.trustnet.com/Tools/Equities/FactsheetPDF.aspx?code=HMWO

Its Dividend has fluctuated quite a bit since 2013, starting at 30.26 cents, rising to 39.11 in 2017 and only 28.14 last year. But if someone smoothed this income volatility by spending from cash or bonds, they would still have their equity capital and maybe this would be better than having to have sold HMWO last March when it was down 25% in a month?

@Whettam – I’d agree GENERALLY not as volatile as asset prices (sorry for shouting) though the figures you quote show an (exceptional?) fall in dividends of about 25% 😉

But I would hope/plan to have 20%-30% or so in bonds to sell instead. At (puts finger in air) 3.5% WR that 6-8 years or so with no real return on bonds. I hope that any share price recovery should be under way by then though a 1966 scenario gives me the creeps. As that begins to run out, can I reduce spending? Stack shelves for 10 hours a week? (Actually, my SP should kick in in 12 years so I’d have an adequate floor.)

@Brod I did not give every years data, but yes the dividend of HMWO dropped 25% over three years, the share price did the same in a month 😉

I wanted a passive example for a Monevator comment 🙂 but I could have picked lots of other active Investment Trusts e.g. Bankers in share price performance terms a closet all world tracker, but 21 years of dividend growth, no drops since 2000 or BlackRock Smaller Co’s 13 years of dividend growth.

I’m NOT (sorry 🙂 saying Dividends are the total solution, but I think their unique characteristics, make them worthy considering for deaccumulation.

@Al Cam
re Zwecher’s book – £61.75 for a Kindle copy? No thanks! I’ll see if the library can source it for me otherwise I’ll rely on Retirement Cafre et al. Cheers anyway :=)

@Whettam (#88)
Re @Naeclues approach:
I agree with you and IIRC @Naeclue and I discussed it at some length a few months back on Monevator as it was emerging. It has probably matured a bit since then but what I particularly liked was:
a) it did not rely on bonds;
b) it’s elegance; and
c) that it ticks the boxes (reversibility, etc) that I mentioned at #63 above.

@TA, if you make a return on reinvested dividends above the cost of trading those dividends + the return on cash, then you are winning. That may well be the case if you don’t run out of bonds as you will always be selling shares for more than you paid.

The difficulty is if you run out of bonds. Then you become a forced seller of shares and more likely to lose out on the dividend trading.

Spending the dividends clearly reduces the risk of running out of bonds. Alternatively, that means you can take the same risk of running out, but with fewer bonds, ie you can up the initial equity allocation. Upping the initial equity allocation increases the expected growth.

Without doing a lot of work, which may not produce a satisfactory answer, it is not obvious to me whether reinvesting dividends is likely to be better or not. I would not feel comfortable holding 6 years of bonds/cash (I started with 5) if I was not also prepared to spend the dividends.

What we do know is that, all else being equal, trading increases costs, so in the absence of further evidence, it seems better to me not to reinvest the dividends.

@Al Cam, yes I do have unsheltered holdings and none are accumulating funds, but that would not sway my decision on whether to reinvest dividends or not. There are still costs involved with dividend reinvestment in accumulating funds.

@TA (#101):
Assuming this Q was to me; I was not looking at any tax advantage, just eliminating two sets of explicit fees to generate income (via the re-investment by the individual route) when income would be available with no associated need to pay fees (via the spend dividends by the individual route).
However, as @Naeclue points out at #102 this may be a misunderstanding on my side as presumably similar [what I assume must be implicit] costs are associated with reinvesting via ACC units.
I must admit though that I am struggling to believe that the costs (explicit vs implicit) of the two routes (re-investment by an individual vs re-investment by the fund) are actually that similar due to the relative scale of the two operations.

@TA, dividends in accumulating funds and ETFs have to be reinvested by the fund managers. They will pay fees, duties and potentially suffer from a spread, all admittedly at lower cost than retail investors, but there is still a cost. When taking money out there may also be costs, such as dealing fees, spreads or swing pricing with OEICs.

The trouble with changing strategy when you have cognitive decline, is that as your cognitive decline takes hold you actually get more confident that you know what your doing, not less! so you will never realise your going to end up with nothing! what a typically human dilemma ‍♂️

@TA, you will not see any tracking difference between income and accumulation units of the same fund. The reason is that the calculation of the total return from income units assumes identical reinvestment to that of the accumulation units. In practice there may be differences due to timing of reinvestment and extra costs payable on reinvestment. For example, some brokers charge to reinvest dividends.

If you are going to be reinvesting dividends then ACC units make perfect sense as reinvestment is likely to be done at the lowest possible cost.

If you are drawing an income, INC units make more sense as you then have to sell a lower value and so lose a smaller amount in the selling. For large index funds and ETFs, we are probably talking about tiny amounts though. Differences in timing are likely to swamp any observable difference.

I see the main benefit of spending dividends as being able to hold more in risk assets than would be the case if dividends were reinvested. If you are paid a dividend it becomes part of the defensive portfolio. Reinvesting the dividend puts it at risk again.

Consider your 60:40 portfolio and 3.8% SWR. At constant SWR and assuming the defensive portfolio kept track with inflation, it would take about 10.5 years to exhaust the pot. If instead you went for a 70:30 pot and you averaged 2% in dividends from the growth portfolio, it would take 12.5 years to exhaust the defensive pot. Even if you only averaged 1%, it would still take 10.7 years to exhaust the pot. There are of course drawbacks. It is clearly going to take longer for the growth portfolio to grow by 20% if you are spending instead of reinvesting the dividends, so defensive pot top-ups can be expected to be less frequent.

@Whettam (#70), @Naeclue (#71), @AlCam (#73).
Sorry for the late reply. Unfortunately still have to do the day job! Thank you very much for your insights.
Just to clarify. Does this mean that if on my 55th birthday for example I have £1million in my pension and I decide to crystallise £200K into a separate pot, I can still pay into the other £800K pension (employer contributions to avoid being classed as recycling) as long as I only take from the tax free part? I am also not at risk of having to pay tax re the LTA even if the total amount exceeds the LTA. I would then intend to drawdown from about age 59 and the value would subsequently go below the LTA by the time I reach 75. Hope this makes sense. Thanks

@Whettam:
I have to say I admire your admiration for pension wise.
What is this based on?

@Naeclue:
Re: “Crystallising everything eliminates the risk that your existing pot will grow to exceed the LTA”
I assume you are ignoring the second test at age 75 for brevity.

@TI Should one factor in the decreasing number of companies that are choosing not to be listed on public markets? It would seem that us great unwashed are being excluded more and more from the good investments.
“The number of companies listed on the London Stock Exchange with shares available to the general public has fallen by half since its peak in 1975; and almost half of those remaining listings are now on AIM—the secondary market that was created in 1995 with much weaker supervision than that applied to the main market. The US has experienced a similar decline in the scope of public markets.
Over the last two decades, less money has been raised on the stock markets of Britain and the US than has been taken out through acquisitions for cash and “share buybacks”—the process by which a company uses its cash to purchase its own shares on the stock market.”
https://www.prospectmagazine.co.uk/magazine/plc-corporation-private-equity-companies

@ Naeclue – this is interesting and I’d just like to kicking the tyres of this a bit more if that’s OK? No worries if I’ve tried your patience long enough.

Take the 70:30 example again. This time, you live off the bonds and reinvest the dividends in your equities and allow them to compound.

When your bonds run dry you sell the reinvested dividend portion of your portfolio that’s been compounding in the background. You sell that into bonds and live off that money until the bonds run dry again. At which point, you sell the dividend growth portion again. How does that look?

Re: acc vs inc fund costs. Vanguard publish a list of costs incurred by their acc and inc funds here:

https://www.vanguardinvestor.co.uk/content/documents/legal/vanguard-full-fund-costs-and-charges.pdf

The costs are identical across all items. That includes transaction costs where Vanguard report broker commissions, spreads etc.

We could get into the annual reports for a belt and braces check.

@TA, An interesting option. You could do as you say and reinvest dividends into another pot then draw them if you run out of bonds. Whether you would want to invest all dividends back into the growth pot is a matter of choice. I only keep 6 years of safe assets (cash at present), and my preference would be to invest the dividends in safe assets. If you are holding bonds that will last longer than that then perhaps reinvesting some or all back into risk assets makes sense.

Or maybe a hybrid approach of some sort, eg reinvest into growth, but if you get down to only 4 years of bonds, switch to spending dividends (or reinvesting in safe assets). Or have some some kind of sliding scale.

I quite like this approach, but it would not work for us as our strategy says if we hold more than 6 years in cash the excess is given away (our strategy incorporates beneficiaries interests).

Regarding fund charges and other costs, yes they are the same for ACC and INC. Not really sure what the question is, but if you are paid a dividend then at that point the dividend is no longer subject to costs and risk. If you choose to reinvest, then the dividends become subject to costs and risk. This happens irrespective of whether you hold INC or ACC units. If reinvesting is the intention, then ACC is likely the most efficient choice, but if you want to spend the dividend, INC is likely to be more efficient than selling ACC units as it avoids the dividend getting round tripped.

@Naeclue (#122):

Looking again at your numbers I assume to get to an overall 30% hit you are incorporating some combination of: a) annual tax free allowance and b) that not all of the SIPP is subject to the LTA (ie when you fully crystallised at 55 there was still some LTA remaining). Otherwise, I would have expected the overall hit to be 40%.
Is that about correct?
And if so I guess you will also have had to make some assumption about future growth and does the calculation allow for the drag introduced by this weeks budgets freezing of the LTA too?

@Al Cam, I forgot to mention I had assumed no withdrawals between crystallisation and BCE5A. Another extreme case.

@Naeclue (#125):

As I see it:
a) as presented (noting #126, etc) the hit on the growth is 40%; but
b) if the money were in an ISA the hit on the growth would be 0%.
Only remaining issue is any net hit in transforming the SIPP to an ISA.

IMO, the best way to minimise the LTA hit is to avoid it – if you can!

@gmo (#133)
Lots of details there, and I think I generally agree with most of it.
There are a couple of other key points I can think of off the top of my head – neither, or one, of which may not actually apply to you and/or me.
These are:
a) presence (but possibly not yet in payment) of a DB scheme pension, which combined with SP pushes taxable income close to (or even closer, due to latest budget) HRT;
b) significant GIA holdings

IMO none of the regimens you talk about seem to last terribly long these days – so there is inevitably an element of speculation required. And, win or lose, these are absolutely not the worst problems to be facing!

I think the observations on smoothing are fair enough if like ERN you want to model and process monthly. This was not a consideration for me as I would be happy to rebalance and DC sort drawdown approx annually. Living off your money book goes on to examine combining PH with variable income methods which ERN indicates would be interesting to explore further. In the end all the differences examined are so small as to be inside the error bar (of total risk including speculative risk vs backtested known market risk). For me these methods perform very close to “the same” for all practical purposes.

Additionally the observations mildly discounting past good performance of PH when bonds did well and equities lay untouched are to my mind not as well considered as the other analysis. PH (smoothed or not) backtests for all past cohorts vs any other method and gives a particular result. Once you start editing the known market risk scenario to take out bits you don’t like you are doing something different. Modelling a different market. McClung also does this in later chapters to create a stress market worse than the known history series.

The observation that we are unlikely to see a bond boom (rising interest rates when/if inflation is acknowledged as ticking up past targets (and central bank buffers) are a counter pressure on capital value vs any rise in yield. But the future is unpredictable as is often the case.

Just a follow up on my experience with updating McClung’s prime harvesting (PH) pst #55. And relevant to preceding posts from @gmo.

I had an attempt at revising my original PH spreadsheet (initiated 4 years back) to account for my changed portfolio construction – I’ve significantly more equity exposure now. The spreadsheet from McClung’s website is slightly updated to a version 2. However, it is tricky (impossible?) to find the correct information to enter into the initial withdrawal worksheet. Key stats are required from the US Fed Stats dept. – (the link in the spreadhseet is inaccurate but I found the relavant document here: https://www.federalreserve.gov/releases/z1/20210610/z1.pdf
If anyone knows the required stats maybe pist them in this thread. I used LM883164115 and LM883164105 – pure guesswork).
The alternative is to skip this step and put in your best guess at a starting withdrawal rate, but this seems to defeat a lot of the point of using the spreadsheet.

Further data is required on Emerging markets and Developed World ex-US CAPE but these are behind a paywall now, and not easily found on the internet in my hands.

So I’ve given up on this attempted revision for the time being. Maybe a table with a few basics stats like these could be maintained on Monevator?

My difficulty updating my PH spreadsheet highlights the longevity risk of relying on weakly supported or unsupported software and calculators. Maybe some of the larger online FI communities have a greater chance of maintaining resources like these over a period compatible with the length of a retirement. At least ERNs calculators rely on easier to obtain information that has a greater chance of being maintained long term.

@gmo #136 makes a good point about how many of the differences between retirement spending models are in the noise, and much of this noise is ‘behavioural’, certainly in my case. I have found comparing model outputs reassuring during the first 6 years of retirement though, as they triangulate around pretty much the same results using different approaches.

I’m quite happy to play around with all these calculators and spreadsheets, but I don’t know for how much longer. And I guess there are quite a lot of folk who are not. I suppose this is a simple 4% SWR rule plus ‘floor and upside’ monitoring when you make an annual withdrawal are appealing. But they are also risky and probably inefficient.

Maybe a post on how simple can you take retirement planning and still be ‘safe’ is worth a thought. Of course, using most of your portfolio to buy an annuity is the gold standard for safety. Maybe that’s the way to go long-term.

I was not aware that Meb Faber had sugested trend following as a useful retirement strategy. However, there are several academic papers by Andrew Claire and collegues at Bayes Business School (formerly Cass) that argue this case, and also that it is a good defence against sequence of returns risk. Here’s an example “Can sustainable withdrawal rates be enhanced by trend
following?” https://onlinelibrary.wiley.com/doi/epdf/10.1002/ijfe.1774

I’ve not looked into this carefully for my own situation as one thing that is not considered in their models are the tax implications of switching your entire portfolio in and out of equities something like once or twice a year. Perhaps if the strategy was applied only to the tax sheltered part of your portfolio it might be worth considering, otherwise the capital gains tax would be exessive. Having said that, with CGT relief of £12,300 you can sell up to £164,000 tax free in the tax year assuming 7.5% return (£12,300/7.5% I think…). Plus I guess if the market is sinking you’d be selling on the down trend, so losing some of your returns that year.

It’s all a bit active sounding really… OTOH I was coincidentally redistibuting my portfolio between several low cost brokers in early 2020 when the covid flash crash occurred and had a lot of it in cash, so did well in the end by pure luck. But it does show if you can switch into cash at the right time it may work out to your advantage – but it is surely a gamble as an investing strategy.

IanH – what a great link, thanks.
It references the Faber paper and uses the 200 day simple moving average, which is much less TA-like than some of the other trend strategies and looks much safer and stress reducing.
It also address global equity and multi-asset portfolios.
Shows that the strategy works very well for bumping up the SWR.
Think I am going to use it for my SIPP and ISA accounts.

Another good link – I haven’t spent much time on that site – I got a bit of a windfall last xmas when the company I worked for was bought out and have since been trying to figure out how to ‘safely’ invest this, and my consolidated SIPP funds, after figuring out that I might actually be able to retire early if I avoid doing anything too stupid.
I hadn’t really paid attention to investing so am trying to bring myself up to speed now without getting overwhelmed/misdirected by the amount of information out there in the FIRE movement.
So far I’ve read ‘The Ivy Portfolio’, ‘Shareholder yield’, ‘Global Value’ and ‘Global Asset Allocation’ of Mebs books, several of which you can get for free from his site, or £2 as Amazon Kindle.
I’ve enjoyed them all, quick reads and good summaries of stuff I had read elsewhere. And some good ideas.
With bonds at the end of their bull run, inflation coming and valuations high I realise I need a smarter portfolio than just a 60/40, and get to the point that I have a decent level of confidence in what I am doing.
I did mess around buying and selling Ethereum to get some ‘skin in the game’ and experience the difficulties of fighting behavioural baises, which was useful experience.
I am not worried about under performing the market so much, more about decent returns to at least keep up with inflation, and safety. The fact the gains are sometime more is a side benefit, I’m more interested in the smoothing effect of this strategy – and mechanically following the triggers could remove the anxiety caused by the behavioural finance issues.
Plus I am a programmer so was thinking that I could automate a lot of the trigger monitoring to reduce the management burden – choosing different indexes for different parts of the portfolio.

I like Living Off Your Money’s spreadsheet because it clearly shows inflation adjustments, I’m starting the book this weekend. Currently trying to figure out if ERN’s Cape Based Rule requires an adjustment for inflation after the Target Withdrawal ($/month) is calculated, but I can’t find a definitive answer on that.