Growth assumptions in your models/spreadsheets

Hoenir

optoutDB said:

The values being used for growth in this thread match what I've been using up to now, but I would call them my pessimistic scenario.   All the Youtube (not)financial advisors seem to go for 100% equities until near retirement and assume 5% growth or so. So maybe I expected some more like that. But I get why caution would be the standard.

The failure point is simple. Growth isn't linear. Not just a question of compunding. The real art is to lose less Beta in a downturn. Boils down to basic maths. If your portfolio drops 20% in value in a correction. Then it has to return 25% in order to breakeven. 

If your portfolio only drops 10%. Then it has to return 11% in order to breakeven. 

Diversfication isn't a figment of someone's imagination. Truly does a serve purpose. No one can say when the insurance will come in handy though. 

optoutDB

michaels said:

So what would be your plan, perhaps some sort of tranches of risk, so annuity (plus DB/SP) for a floor income, low average return lower risk for the next tranche and then high average return high volatility for the 'fun' tranche?

What sort of risk level would you want for each tranche?  80% success for the low risk tranche and 50% for the fun tranche?  That still means that 1 in 5 retirements are spent at floor income.  All sounds like casino retirement to me TBH.

ooh!

your 1st paragraph summarises my strategy more concisely than I could.

I don't see why 1 in 5 retirements would be spent at floor though. Is that sequence analysis with a dumb expenditure model?

OldScientist

optoutDB said:

Just catching up with this thread,

The values being used for growth in this thread match what I've been using up to now, but I would call them my pessimistic scenario.   All the Youtube (not)financial advisors seem to go for 100% equities until near retirement and assume 5% growth or so. So maybe I expected some more like that. But I get why caution would be the standard.

I would prefer realism though. and I think the caution should come in to where you put your wealth.

I wouldn't mind getting hold of the historic data by asset class, so I can go though the process of doing sequence analysis. Running all previous time spans does seem a reasonable way, compared to running pessimistic values.

BUT,  any model with non-adaptive expenditure looks near worthless to me. The scenario that anyone would put eg 80% into equites and then spend themselves into oblivion through a market crash is not worth looking at, and that's what your X% chance of failure due to sequencing is modelling. I've only looked at one online DC pot model, so maybe there are better ones.

Even if my equities go to zero over the next 10years, I won't run out of money in 30 years time because I will spend less.  

The target of my modelling has always been how much can I spend next year given that I want to maximise the total spend before I die, but with a front loaded profile. It's easy to do low spend and assume growth to get a big pot when you are 80.

Asset returns (and inflation) can be found at macrohistory.net (individual countries - you'll have to construct your own international GBP based index - I've used a simple 50/50 mix of UK/US as well as domestic).

IMV, the constant inflation adjusted withdrawals ('SWR') approach is useful as a guide for pre-retirement planning but would require a lot of nerve to operate in a difficult market.

There are a load of adaptive models - the simplest is a constant percentage of portfolio (e.g., take 5% of the portfolio value each year) while there are variants where the percentage changes with time (e.g., VPW or ABW on bogleheads or the US RMD approach which is based on longevity) but all of these will provide volatility in income equal to the volatility of the market but have no chance of prematurely running out of money (although income can become small).

Then there are hybrid methods that mix constant inflation adjusted withdrawals with percentage of portfolio methods (e.g. Carlson's endowment, Vanguard dynamic withdrawals, Guyton-Klinger, etc.). Compared to the percentage of portfolio methods, these all reduce the volatility in income at the expense of increasing the chance of premature portfolio exhaustion.

IMV, the first question is how much volatility in income can you stand? The answer to this depends on sources of guaranteed income (state pension, DB pension, RPI annuities) compared to core expenditure (i.e., expenditure that you cannot or are extremely unwilling to forego). The amount of desirable volatility may also vary with time as different sources of guaranteed income come online.

For example, for a couple each with a full state pension (i.e. £23k) and core expenditure around £23k are not going to care that much about income volatility from their portfolio so could pretty well go for any withdrawal approach they like.

Conversely, a similar couple whose core expenditure is £30k are going to want to choose a withdrawal strategy that is likely to give at least £7k per year (or they just buy an RPI annuity or construct an ILG ladder to provide that £7k).

MK62

michaels said:

MK62 said:

Bostonerimus1 said:

MK62 said:

While running out of money is potentially the ultimate failure (although I doubt many would stick rigidly to a 30yr withdrawal plan if, after say 20 years it looked like running out), it's not the only one..........some might see that living off a lower income than you needed to in your prime retirement years is also a fail, and statistically, that's the most likely outcome from following the SWR strategy alone.....there is no rewind button on your life.
In the end, it's a tradeoff of probabilities, so perhaps for some, a more balanced approach is warranted.....both in asset allocation and drawdown strategy.....or else scratch the drawdown only approach and buy an annuity with all or part of your portfolio.

I don't necessarily agree that spending more money equates to retirement success. Of my regrets none of them are about spending money.

That's OK......everyone is different, and I did say "for some".

There are degrees of success though........in my view, while all 30yr plans which end with a positive balance can be viewed as a success in terms of not running out of money, they won't all have the same degree of success in terms of maximising income during that 30yr retirement........so I suppose it depends on your priorities. There's also the question of income stability to consider too........though again, this might be less of a priority for some than others.

So what would be your plan, perhaps some sort of tranches of risk, so annuity (plus DB/SP) for a floor income, low average return lower risk for the next tranche and then high average return high volatility for the 'fun' tranche?

What sort of risk level would you want for each tranche?  80% success for the low risk tranche and 50% for the fun tranche?  That still means that 1 in 5 retirements are spent at floor income.  All sounds like casino retirement to me TBH.

In drawdown, for a constant IL withdrawal, such as the SWR strategy, each has to decide on a historic success rate for not running out of money vs the level of income that will provide........that might be 100% historic success at 3% withdrawal rate, 85% at 4% WR, 70% at 5% WR......and so on. However, historic success may not mean future success, but for those who insist on 100% historic success, currently, if that will only return 3%, why not consider an IL annuity (at 60) paying iro 4.2% (single life) or iro 3.8% (joint life), if not for the whole, at least part of the portfolio?

However, that said, if drawdown is the chosen path, while SWR puts you in the ballpark at the start, it's not a strategy I would follow for a full 30 year retirement plan. The first thing I would ask myself is am I starting, for example, year 9 of a 30yr plan, or year 1 of a new 21yr plan?..........the former will mean your income for this year was decided 9 years ago, regardless of what's happened since......
So, for me, a drawdown only strategy would be annual re-evaluation, with variable income supported by a buffer. There are, of course, many ways to implement that, a lot of which will simply come down to personal preference......

As to your idea of "tranches", most of us are already doing that with asset allocation......and as for success rates, have you considered that if you base your withdrawal plan solely on the worst case of the last 50 thirty year periods, then 49 of them will be better than that worst case....98%. True, to some degree it's "playing the odds", but then so is investment itself - if you aren't prepared for that, then perhaps an annuity is the way to go.

In the end nobody knows today what the best drawdown plan will have been.......just as nobody knows what the best portfolio will have been......each has to evaluate and trade off the risks and probabilities. If I retired at 60 though, by the time I was 75, I'm fairly sure I would have some feelings of regret if it turned out I could have lived those prime retirement years on a higher income from my chosen portfolio, and I hadn't taken steps towards that......others might not feel that way though.

OldScientist

Further to my earlier post (off topic, but the OP did bring up adaptive withdrawals)

The following graph shows (top panel) the real withdrawal (expressed as a percentage of the original portfolio) and (lower panel) the real portfolio value as a function of time for a UK retirement starting in 1937* (one of the worst historical cases) for different amounts of income flexibility, F (implemented as a simple mix of SWR and percentage of portfolio with a starting withdrawal of 3.5%), where F=0 is no flexibility, i.e., constant inflation adjusted, F=20 is 80% SWR and 20% percentage of portfolio, F=100 is fully flexible, etc.).



The above graph illustrates the compromise between low income volatility with possibility of portfolio exhaustion (and consequently zero income) and high income volatility and reduced possibility of portfolio exhaustion.

There is no 'correct' or 'optimal' solution here - only solutions that might be acceptable. Adding in guaranteed flooring (e.g., state pension, DB pensions, RPI annuities, and ILG ladders) will make the volatility even of the pure percentage of portfolio look a little better, e.g. the following graph shows the total income using the pure percentage of portfolio after adding guaranteed income of 0%, 2%, and 4% of the original portfolio. The absolute variation is the same, but the percentage change in income is less as the floor increases.



* 60% UK stocks, 40% UK long bonds, UK inflation (asset returns and inflation from macrohistory.net)

michaels

I find that final chart a really useful visualisation (just checking my understanding, that is the pale blue line of chart 1 plus 0%/2%/4%?)

michaels

MK62 said:

michaels said:

MK62 said:

Bostonerimus1 said:

MK62 said:

While running out of money is potentially the ultimate failure (although I doubt many would stick rigidly to a 30yr withdrawal plan if, after say 20 years it looked like running out), it's not the only one..........some might see that living off a lower income than you needed to in your prime retirement years is also a fail, and statistically, that's the most likely outcome from following the SWR strategy alone.....there is no rewind button on your life.
In the end, it's a tradeoff of probabilities, so perhaps for some, a more balanced approach is warranted.....both in asset allocation and drawdown strategy.....or else scratch the drawdown only approach and buy an annuity with all or part of your portfolio.

I don't necessarily agree that spending more money equates to retirement success. Of my regrets none of them are about spending money.

That's OK......everyone is different, and I did say "for some".

There are degrees of success though........in my view, while all 30yr plans which end with a positive balance can be viewed as a success in terms of not running out of money, they won't all have the same degree of success in terms of maximising income during that 30yr retirement........so I suppose it depends on your priorities. There's also the question of income stability to consider too........though again, this might be less of a priority for some than others.

So what would be your plan, perhaps some sort of tranches of risk, so annuity (plus DB/SP) for a floor income, low average return lower risk for the next tranche and then high average return high volatility for the 'fun' tranche?

What sort of risk level would you want for each tranche?  80% success for the low risk tranche and 50% for the fun tranche?  That still means that 1 in 5 retirements are spent at floor income.  All sounds like casino retirement to me TBH.

In drawdown, for a constant IL withdrawal, such as the SWR strategy, each has to decide on a historic success rate for not running out of money vs the level of income that will provide........that might be 100% historic success at 3% withdrawal rate, 85% at 4% WR, 70% at 5% WR......and so on. However, historic success may not mean future success, but for those who insist on 100% historic success, currently, if that will only return 3%, why not consider an IL annuity (at 60) paying iro 4.2% (single life) or iro 3.8% (joint life), if not for the whole, at least part of the portfolio?

However, that said, if drawdown is the chosen path, while SWR puts you in the ballpark at the start, it's not a strategy I would follow for a full 30 year retirement plan. The first thing I would ask myself is am I starting, for example, year 9 of a 30yr plan, or year 1 of a new 21yr plan?..........the former will mean your income for this year was decided 9 years ago, regardless of what's happened since......
So, for me, a drawdown only strategy would be annual re-evaluation, with variable income supported by a buffer. There are, of course, many ways to implement that, a lot of which will simply come down to personal preference......

As to your idea of "tranches", most of us are already doing that with asset allocation......and as for success rates, have you considered that if you base your withdrawal plan solely on the worst case of the last 50 thirty year periods, then 49 of them will be better than that worst case....98%. True, to some degree it's "playing the odds", but then so is investment itself - if you aren't prepared for that, then perhaps an annuity is the way to go.

In the end nobody knows today what the best drawdown plan will have been.......just as nobody knows what the best portfolio will have been......each has to evaluate and trade off the risks and probabilities. If I retired at 60 though, by the time I was 75, I'm fairly sure I would have some feelings of regret if it turned out I could have lived those prime retirement years on a higher income from my chosen portfolio, and I hadn't taken steps towards that......others might not feel that way though.

1) Re your underspend regrets - isn't that a bit like saying I wish I hadn't bought house insurance for the last 15 years as I never claimed on it?

2) I certainly think there is room for an annuity when rates exceed historic SWR (or whatever other 'worst case return scenario you use in your modelling) with the caveat at at least partially one can justify the fact that taking a conservative approach to spending does not result in the money being 'lost' but actually simply going to future generations - which is not the case with an annuity.

optoutDB

Thanks for the inputs, I'm glad to see all these ideas are out there !  My fault for lack of research, I tend to work things out myself from first principles.

I'm sure when I get a moment I will get that historic data and do some tests with it.

I realised in bed today (yep it was on my mind) that when I glanced at the SWR model notes it mentioned rebalancing. At the time it didn't stand out, but if the model rebalances the allocation every year and eg keeps you at eg 75% equities through a long term equities downtrend, then that's going to be bad. I don't have rebalancing in my model so in my head 100% of initial allocation is the most any asset class can lose. 


I'm currently 10 years away from SP, and thinking about what lifestyle business to try next. So I've got a wide range of potential earnings to consider on top of everything else, the range is probably zero to doubling my wealth. I should be working on those businesses rather than my forecasts, but I needed to check that drawing my pension now wasn't a terrible idea and then I got sucked in by the numbers. 

Thanks everyone. I'll get back to work now.

QrizB

michaels said:

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

That's pretty much the current situation. The UK SWR is something like 3.5% but a single-life RPI annuity is yielding something like 4.8% for a 65-year-old retiree.

So if you want £14k a year, you can save £400k and draw down. Or you can spend £300k on an annuity and still have £100k in the bank.

Hoenir

QrizB said:

michaels said:

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

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

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