How much of an ARF should be in equities?

sarenco finding it a bit difficult to follow your figures. 150K is 15 years residual expenses? I take this to mean that you target to live on 10k + State Pension + 3k (other DB). Or in other words your target pension income from your current situation is 10k p.a.
Oops - error corrected above, the "comfortable lifestyle" target should of course have read €25kpa.

On the more substantive point, I very much accept that a rising allocation to equities in retirement is very unconventional but I think it strikes a reasonable balance in trying to address sequence of return and inflation risk. I certainly haven't firmed up my thinking on this (or really any!) aspect of my retirement plan.
 
Oops - error corrected above, the "comfortable lifestyle" target should of course have read €25kpa.

On the more substantive point, I very much accept that a rising allocation to equities in retirement is very unconventional but I think it strikes a reasonable balance in trying to address sequence of return and inflation risk. I certainly haven't firmed up my thinking on this (or really any!) aspect of my retirement plan.
I'm being a bit slow here. 150k is 15 years of "residual expenses". That suggests 10k p.a. is your requirement for these expenses. But an annuity would give you 24k p.a. safely and comfortably meeting your requirement.
 
Sorry Duke, I'm obviously not expressing myself very clearly.

The €10k of residual expenses is simply an input in my home brew formula for determining an appropriate initial asset allocation for my hypothetical ARF. I would still take the minimum annual drawdown from the ARF, which I hope would always be significantly higher than €10k pa.

Hopefully that makes sense.
 
Sorry Duke, I'm obviously not expressing myself very clearly.

The €10k of residual expenses is simply an input in my home brew formula for determining an appropriate initial asset allocation for my hypothetical ARF. I would still take the minimum annual drawdown from the ARF, which I hope would always be significantly higher than €10k pa.

Hopefully that makes sense.
no, what is your requirement for residual expenses post retirement? If it is only 10k then 600k retirement fund at 60x protects you from the three main risks: longevity, sequence of returns and inflation. It means that either you have over provided for your retirement or that you plan an overly monastic lifestyle in your golden years
 
Last edited:
I think we're talking at cross-purposes.

My residual expenses (as I've defined it) is what I need (to live a comfortable life in retirement). But I'm hoping that the stock market will give me what I want (a retirement full of travel, fine wine, etc.).

I'm very clear what risks I would be protecting myself against by simply buying an annuity.

I'm also clear about the potential upside that I (and my estate) would be foregoing.

So, I'm trying to land on a "goldilocks" allocation within an ARF.

I may well annuitise a portion (or even all) of my retirement savings at some point. But I wouldn't do so immediately if I was retiring today.

Would you?

 
Just to spell it out clearly because the numbers probably are similar for others


3k DB
12k state pension
10k from pension plan

Next to understand what you're saying, 150k cash drawdown immediate. That covers 15 years@10k per annum

Meanwhile the 450k is invested and draws down at 6% per year

Face value, after 15 years that has drawn down .06 x 15 x 450k = 405k

I'm probably getting it all wrong
 
Sarenco,

I’m not convinced that’ll be enough for you to live on as you’ll be hosting me for booze-fuelled debates when we’re too old for this malarky.

Gordon
 
Ok, I think I know where sarenco is coming from. He (apologies if she) has thrown me off kilter with this reference to "residual expenses". So getting back to OP and reframing the question.

"What is the best ARF strategy for achieving a target subject to an acceptable level of risk?"

Still very woolly. So to get some focus let us consider a 65 year old retiree and make something a bit more concrete of this aspiration.

"For a 65 year old what ARF strategy maximises average level withdrawals over 30 years subject to a safety net that the risk of not lasting 20 years is less than 5%?"

Note that without the safety net caveat the answer is unquestionably 100% equities. But that introduces sequence of return risk. Putting in a safety net raises the possibility that optimality will involve less than 100% equity exposure.

Note also that every aspect of the reframed question is personal, the 30 year target, the level withdrawals, the 20 year and 5% risk calibration etc. However in breaking the elephant up into these pieces Mr/s Normal might be able to give plausible answers where s/he would be staring into a fog faced with the generic question posed at the start of this post.

Finally I come to sarenco. As I understand it the sarenco conjecture goes as follows:
sarenco conjecture said:
The optimum strategy for achieving the ARF objective involves a mix of cash* and equities and in particular withdrawal of the cash before decumulating the equities.

* For simplicity I assume only two asset classes - equities and cash.
If this is a misinterpretation of sarenco let the conjecture be called the Marmalade Conjecture.


As sarenco points out the conjecture leads to the counter-intuitive notion that equity exposure increases during the decumulation phase. Also it is not a priori clear that the conjecture is correct. The strategy involves exposure to equities for longer and this cuts two ways - it increases average expectations but it also increases risk.

If the conjecture is correct the supplementary is what is the cash/equity split? sarenco posits 15 years of residual expenses in cash and the rest in equities.

To answer these questions needs the use of generally accepted stochastic methods. I am on the case.:cool:
 
Last edited:
These are the preliminary results from my investigation into the sarenco conjecture
Assumptions: Equities earn on average 6% p.a. but with an annual volatility of 20% p.a.; cash earns 0.5% p.a.
ARF Objective: To maximise annual level withdrawals over 30 years subject to a safety net of 5% chance of bomb out before 20 years

The sarenco strategy (Cash out first/Equities out last) gets optimised with a 85% initial cash allocation and the level withdrawal is 5%

First thing we note is that the 5% safety net is dominating, giving rise to a high cash allocation; nonetheless we are here to test the sarenco conjecture so we optimise the alternative strategy of withdrawing equal percentages of cash and equity each year. The results are that this strategy gets optimised with a 71% cash allocation and provides an annual withdrawal of 4.8%. Slightly less than the sarenco optimum of 5% p.a. and so tentative evidence that there is some validity to the conjecture.

I then relaxed the safety to a 20% acceptance of bomb out at 20 years and got the following results:

sarenco optimisation kicks in at 75% initial cash allocation and provides a level withdrawal of 5.5% p.a.
constant allocation optimisation kicks in at 60% cash allocation and provides a level withdrawal of 5.2% p.a.

Further confirmation that there is at least some validity to the sarenco conjecture.

But of course this can't be the basis for the Holy Grail, Bob2018. Any system that has 90 year olds 100% in equities would be laughed out of court. It may be a basis for a dynamic strategy that steadily adjusts the safety net. It has the merits that it appears to tap the potential for equity outperformance with a lesser opening exposure and without the immediate need to sell equities.

Caveats: this is a hugely simplified study to try and assess the validity of the conjecture. It ignores all sorts of practical considerations such as tax, deemed distribution, other asset classes, costs, inflation etc.
 
Last edited:
Any system that has 90 year olds 100% in equities would be laughed out of court. It may be a basis for a dynamic strategy that steadily adjusts the safety net. It has the merits that it appears to tap the potential for equity outperformance with a lesser opening exposure and without the immediate need to sell equities.
On a point of detail Duke, I suggested an upper ceiling of, perhaps, 80% in equities.

However, the gist of my argument is that it is less dangerous for a late retiree to have an equity heavy allocation than it is for an early retiree because of sequence of returns risk.
 
But of course this can't be the basis for the Holy Grail, Bob2018.

WHAT? WHAT'S THIS NONSENSE? You haven't found the Holy Grail? What the hell have you been doing all day?

I feel very "grown-up" having on-line discussions about "generally accepted stochastic methods"! If only I could follow it all! (in fairness, I think your writing style helps!) How can we learn more about your analysis (like what is the Holy Grail)?

I read this today. It seemed interesting to me but I'd be interested in what the experts think.

https://www.7im.co.uk/-/media/files/brochure/decumulation-discussion-paper.pdf
 
WHAT? WHAT'S THIS NONSENSE? You haven't found the Holy Grail? What the hell have you been doing all day?
Well think about it Bob. If I had found the HG today would I announce it on AAM or would I be straight down to the patents office? So you will never know:D
 
I know that deep down and very well hidden the Duke is a decent skin - so he will in time, like the financial advisory community, share his full wisdom with us all! ;) On a serious note Duke, can you use the "upload a file" facility to share either your previous or current spreadsheet?

The point I'd like to make now is a bit peripheral to Sarenco's question but probably worth mentioning. If I understand Sarenco's situation correctly, he "has" an ARF of €600k and "needs" €10k p.a. to cover required expenses.

On the one hand, the "Authorities" are trying to promote pension provision (e.g. the introduction of mandatory PRSAs) but on the other hand, they are placing unwelcome hurdles on Mr. Normal in the form of imputed distributions. I accept that Mr. Normal does need to take money from his ARF during his decumulation phase. However, it's the almost compulsory requirement to withdraw funds on an annual basis that is problematic.
The problem is that these forced withdrawals increase sequence of returns risks and associated concerns.

In Sarenco's case, with €600k in an ARF, presumably he could take €150k as a tax free lump sum thereby providing him with multiple years of guaranteed income with the number of such guaranteed years being a function of his preferred asset allocation, i.e. investing in cash, more or less, gives him 15 years.

If Sarenco had this security (that his base needs were covered for many years and that he wasn't forced to withdraw funds at "the wrong time"), he would be so much freer to be more expansive in the asset allocation of his residual ARF investment of €450k. In simple terms, sequence of return risk would have a significantly diminished impact if Sarenco was afforded greater freedom in the timing of withdrawals in his decumulation phase.

There is a lot of data showing that the asset allocation of ARFers is too conservative given the probable long-term lifespan of such investments. It is possible that the introduction of the 3% distribution requirement in 2008 (and subsequent increases) has not been of helpful in this regard. Whatever the case, if you were to try to find a worse time to introduce this requirement, you'd be hard pressed to do so!
 
Last edited:
@elacsaplau

In my hypothetical fact pattern, I had already taken my TFLS and used it to pay off my mortgage.
To put some numbers around this, let's say I'm retiring today at 66, with a paid for house and I qualify for the full State (Contributory) Pension. I have DB pension income of €3kpa from an old employer and an ARF with a value of €600k. For simplicity, let's say I have no material savings outside my ARF (I used the tax-free lump sum to pay off my mortgage) and I have no desire to hold an AMRF.

I estimate that I will need around €25kpa to have a comfortable lifestyle. Importantly, €25kpa isn't necessarily what I would spend on my desired or target lifestyle - it's pretty much a floor as to what would be acceptable to me.
 
In Sarenco's case, with €600k in an ARF, presumably he could take €150k as a tax free lump sum thereby providing him with multiple years of guaranteed income with the number of such guaranteed years being a function of his preferred asset allocation, i.e. investing in cash, more or less, gives him 15 years.

If Sarenco had this security (that his base needs were covered for many years and that he wasn't forced to withdraw funds at "the wrong time"), he would be so much freer to be more expansive in the asset allocation of his residual ARF investment of €450k. In simple terms, sequence of return risk would have a significantly diminished impact if Sarenco was afforded greater freedom in the timing of withdrawals in his decumulation phase.

The required withdrawal certainly does imact on a person's investments within an ARF for the reasons you outline. My understanding is that mandatory withdrawals were introduced because the Powers That Be looked at the ARFs that existed up to that point and discovered that there were some whopper ARFs out there - multi-million euro funds - from which no income was being taken at all. Turned out that ARFs were being used as a vehicle by which wealthy business owners could extract millions from their companies through pension schemes, transfer them into ARFs, take no income from the ARFs (because they were minted anyway and didn't need it) and then pass the millions on to their descendants after death. The Powers That Be were not happy. The ARF was invented as an alternative to the annuity - an alternative form of drawing retirement income. So using one to draw NO income was not the original intention. Not to mention the fact that by forcing these lads with whopper ARFs to take even a small percentage income, it would still be a whopper income generating a whopper Income Tax return for the Revenue. So mandatory income was born.

Even though most of us won't have whopper ARFs in retirement, I can still see the point that if the ARF is an alternative to the annuity, then a mandatory income is justified.

In the example you give above where an individual lives on their tax-free lump sum for the early years of their retirement, I suppose one way of avoiding sequence risk is to re-invest the mandatory ARF withdrawals into the same type of investments as the ARF itself. A bit messy and there would be charges and taxes involved.
 
Apologies Sarenco - I misunderstood.

Aaaggh Liam - most of what you've said is indisputably right! and the balance is not wrong! Should I just delete my last post? ;) [......I had thought the dodgy line was teasing the Duke but so far I seem to have gotten away with that one! I suspect that I'm not out of the woods there either...….revenge est un plat mieux servi froid!]

Whilst I have you on the line, as it were, Liam! - anything else to add (solutions wise!) on de subject?! :D
 
Well elac thanks for pointing the way to the Upload File facility. I have duly uploaded the original version with an explanatory sheet added in. Corrections, comments and suggestions for improvement most welcome.

LDF is totally right on the origins and rationale for mandatory withdrawals. But other than its tax consequences it should be of itself irrelevant for asset allocation. I'm surprised some company has not designed an arrangement whereby the mandatory withdrawal has no other impact rather than tax payment i.e. that the underlying assets remain invested exactly as before only now in an exit tax umbrella rather than a pensions vehicle.
 

Attachments

  • ARF Projection.xlsx
    96.4 KB · Views: 230
Back
Top