Ready made portfolios for generating an income in retirement

OldScientist

zagfles said:

Linton said:

zagfles said:

Linton said:

Bostonerimus1 said:

MK62 said:

Bostonerimus1 said:

Triumph13 said:

Bostonerimus1 said:

Triumph13 said:

zagfles said:

Triumph13 said:

I'm a big fan of a cash buffer myself, but not when used for market timing.  When coupled with a fixed withdrawal percentage, I believe (though I admit I haven't run the simulations) that it should have a lot of power to lessen the pain of spending reductions in downturns.  My plan has always been to use it to mitigate 50% of any income reduction - so if markets fall 50% my income taken only falls 25%.  A one year cash buffer would get you through four years of such a 50% drop, which seems like a reasonable reward for the cost of holding it.

More generally though, I suspect most retirees would be best to buy annuities to cover most or all of their baseline spending, making them much better placed to withstand any fluctuations in income from drawdown portfolios.  That would mean that, rather than going for a single 'default portfolio', lots of people should actually be looking to split their funds between two or more different portfolios - a traditional 'lifestyling' one for annuity purchase and an ongoing one for drawdown.  Plus a 'lifestyle to cash' one if they are planning to use the TFLS to pay off their mortgage or bridge a short gap to SP.  You really need to have gone through that thought process before you have any chance of making a sensible decision on which 'default' option to pick, and I worry about the ability of people to get through it without guidance.

If you use a cash buffer to avoid selling equities when the market is "low", then that is market timing.

Or are you saying your withdrawals are a percentage of the current pot value rather a percentage of the initial pot value? That is a very strange way of doing drawdown. If markets rose 50% would you draw 50% more? 

That's exactly what I mean, sell a fixed percentage of the portfolio.  It eliminates SORR, in return for accepting higher volatility on spending.  No market timing involved. The cash buffer then goes some way to smooth the volatility.

If equities were up 50%, I would indeed draw 50% more than planned (or at least convert equities to cash within the wrapper) , all things being equal, on the grounds that if equities were fairly valued I'd have more than enough, if they were overvalued then I'd be better with cash.  I'd then use the excess to increase the cash buffer and potentially spread it over multiple years of increased spending. 

Full disclosure, one of the reasons I can accept such high volatility is that SPs and some little DBs will eventually provide about half our income and are covered by cash in the meantime.

We are really talking about asset allocation with MMF really just being ultrashort bonds and maybe cash being zero duration fixed income. For what it's worth in most scenarios you'll do better the more equities you have up to maybe an 80/20 split when portfolio survivability starts to be impacted by sequence or returns risk early in drawdown. If your time horizon is more than 30 years then high equity percentages really start to pull ahead. So much of retirement asset allocation comes down to the feelings of the retiree rather than the results of the analysis which is why we see cash buffers resulting from a natural fear of volatility. 

You could look at it that way, but that's not the way I designed it or use it.  For me, there is a very clear separation between my fixed income (and associated bridging funds), my drawdown portfolio and my cash as they have different purposes and timescales.  The drawdown portfolio is intended to have a time horizon of more like 70 years than 30 as I want to pass it on to my kids.  It is therefore close to 100% equities.  No portfolio survivability issues because of the fixed percentage approach, just a matter of keeping drawing percentage at or below average real returns over time to preserve the value.   I'm lucky that I can afford to have the income from it be volatile, but a cash buffer to dilute / delay the volatility is still appreciated.

Other people will have different situations and different priorities, but for those lucky enough to have more than their basic needs, I do think a multiple pot approach can be a lot easier to get your head around than trying to think of it as a single asset allocation and planning for how to change that over time.

I understand the desire to dampen volatility with cash. I'm just saying that for most people reserving cash to dampen volatility leads to a greater failure rate ie. there's more chance you'll run out of money by giving up the market gains.

This isn't supported by my own historic simulations, or by online simulators, such as https://www.2020financial.co.uk/pension-drawdown-calculator/#calculator. The failure rate actually goes down a bit rather than up.......the average end balance is lower, granted, but that's a secondary concern, and is perhaps the price for that admittedly marginally lower failure rate. Also, many retirement plans front load early withdrawals pre SP/DB.....right at the time the portfolio is most vulnerable to a poor SOR.

Of course, there are many ways to use cash in a portfolio, and so it's rather difficult to simulate them all......

US studies have shown that historically cash buffers do not improve survivability of a pension pot. It might be different for the UK.

https://bpb-us-w2.wpmucdn.com/sites.udel.edu/dist/a/855/files/2020/08/Sustainable-Withdrawal-Rates.pdf

Anyway, it's a matter of degree as everyone needs cash to spend and I do much the same as I've always done and keep a couple of years spending on hand for emergencies and liquidity.

As a larger point we are talking about managing DC pensions for income here and it's not obvious that people will have enough to support their income needs or the skills to sensibly manage their pension pot. In the US the DC pension model has been around for a couple of generations now and half of US retirees do not have enough to retire, even with the more generous US equivalent of the UK's State Pension. Many people over 65 just keep working, so maybe there needs to be more of a focus on how to use a pension pot for older people who are still working. The self directed DC pension model has failed for many US retirees because of a combination of poor implementation by government, employers and employees. The certain winners have been the financial industry.

Sorry, I have cheated and actually read the article. It uses an Indirect buffer approach and compares the % failures with simply drawing down directly from a single growth portfolio.

The buffer approach is to set up a buffer with sufficient money to pay for n years inflation linked expenditure and replenished each year from growth only if the latter has increased in value. If the buffer is exhausted expenditure is taken directly from the growth investments.

Success is determined by whether the entire portfolio survives for the specified duration of the test period.

I think this buffer approach is much like Prime Harvesting except that in the latter the bond allocation is essentially used as the buffer. Apologies if I have got this wrong.

There are many other ways of running an indirect withdrawal approach. The paper only considers one, so I cannot see it as evidence of any blanket statement about buffering.

The main difference with PH is it's a much longer term "buffer" strategy, if you want to call it that. Even in the worst bear markets it'd likely be over 10 years before you need to sell equities. The term I think is key - short term market timing is a mug's game, but longer term the market is more predictable - after all anyone invested in equities is long term timing the market - ie the assumption that over the long term equities will beat bonds/cash.

It also has researched sell tipping points, ie you don't just sell equities when they've gone up but when they're up 20% real terms. And unlike short term buffer strategies like the one linked above, it appears to generally deliver better results than a fixed asset allocation strategy. More here: Dynamic asset allocation and withdrawal in retirement - Monevator

I've not seen any evidence a short term cash buffer strategy is of any benefit (other than a comfort blanket), obviously the above article doesn't prove conclusively that none exist as the parameters could be tweaked, it's hard to prove a negative (like there's no teapot orbiting the Sun), but if evidence is that "something similar" doesn't work then surely that should be the working assumption unless something comes along to prove otherwise. 

A comfort blanket may well be useful.  In a major equity crash, or even a medium duration down turn what do you do? Cut your drawdown? Or do you carry on drawing down at the same rate regardless, confident from your use of SWR that there is no real risk?  A comfort blanket reduces the need for you to make a difficult decision for a while by which time it may not be necessary.

Personally if I was using a strategy that has been well researched and backtested in different historic market conditions including major equity crashes and prolonged bear markets, a crash on the scale of anything we've seen in the past wouldn't phase me. I'd keep calm and carry on. Why research a strategy and then panic when something happens which the strategy has been proved to cope with.

But if I was using a comfort blanket with no proven real protective effect I might panic and turn the heating down and eat dog food. 

Obviously there's always a risk that the future will be completely different to the past. Who knows, there might be a 30 year bear market a bit like in Japan from 1990, if that were to happen globally no investment portfolio containing significant equities would do well. But some would fare better than others. 

As you know, the SWR strategy has been backtested by using historical data. The initial studies with US data, and subsequent studies with international data (e.g., https://www.advisorperspectives.com/articles/2014/03/04/does-international-diversification-improve-safe-withdrawal-rates )

From that link we know that for a 30 year retirement and a 50/50 allocation to stocks and bonds, the US had a SWR of 3.7%, with New Zealand and Canada having similar values, others had slightly lower values (e.g., Sweden=3.2%, UK=3.1%, Australia=2.9%) and some (e.g., France, Italy, Germany, Austria and Japan) had a SWR below 1% (all countries significantly affected by WWII).

However, the SWR is unknown and unknowable at the start of a retirement. While history is a guide, it is not necessarily going to apply to the future. Taking the Japan example you cite, the following graph shows the MSWR for a 50/50 portfolio for a 30 year retirement in the years since 1950 as a function of start year (returns and inflation from macrohistory.net).




For a retiree in 1989, the historical record would have shown an SWR of just under 5% (the 1962 retirement would have been incomplete, but analysis in 1989 might have suggested the potential for a poor one). The retirements in the early 1970s were only 15 or so years gone so the SWR would have been uncertain in 1989. So, our retiree confidently, but conservatively, uses 4.5% as their withdrawal rate. As it then turns out, the SWR was around 4% and they ran out of money while waiting for the market to recover (see following figure which shows portfolio value in the top panel and income in the bottom panel for a 1989 retiree).

zagfles

OldScientist said:

zagfles said:

Linton said:

zagfles said:

Linton said:

Bostonerimus1 said:

MK62 said:

Bostonerimus1 said:

Triumph13 said:

Bostonerimus1 said:

Triumph13 said:

zagfles said:

Triumph13 said:

I'm a big fan of a cash buffer myself, but not when used for market timing.  When coupled with a fixed withdrawal percentage, I believe (though I admit I haven't run the simulations) that it should have a lot of power to lessen the pain of spending reductions in downturns.  My plan has always been to use it to mitigate 50% of any income reduction - so if markets fall 50% my income taken only falls 25%.  A one year cash buffer would get you through four years of such a 50% drop, which seems like a reasonable reward for the cost of holding it.

More generally though, I suspect most retirees would be best to buy annuities to cover most or all of their baseline spending, making them much better placed to withstand any fluctuations in income from drawdown portfolios.  That would mean that, rather than going for a single 'default portfolio', lots of people should actually be looking to split their funds between two or more different portfolios - a traditional 'lifestyling' one for annuity purchase and an ongoing one for drawdown.  Plus a 'lifestyle to cash' one if they are planning to use the TFLS to pay off their mortgage or bridge a short gap to SP.  You really need to have gone through that thought process before you have any chance of making a sensible decision on which 'default' option to pick, and I worry about the ability of people to get through it without guidance.

If you use a cash buffer to avoid selling equities when the market is "low", then that is market timing.

Or are you saying your withdrawals are a percentage of the current pot value rather a percentage of the initial pot value? That is a very strange way of doing drawdown. If markets rose 50% would you draw 50% more? 

That's exactly what I mean, sell a fixed percentage of the portfolio.  It eliminates SORR, in return for accepting higher volatility on spending.  No market timing involved. The cash buffer then goes some way to smooth the volatility.

If equities were up 50%, I would indeed draw 50% more than planned (or at least convert equities to cash within the wrapper) , all things being equal, on the grounds that if equities were fairly valued I'd have more than enough, if they were overvalued then I'd be better with cash.  I'd then use the excess to increase the cash buffer and potentially spread it over multiple years of increased spending. 

Full disclosure, one of the reasons I can accept such high volatility is that SPs and some little DBs will eventually provide about half our income and are covered by cash in the meantime.

We are really talking about asset allocation with MMF really just being ultrashort bonds and maybe cash being zero duration fixed income. For what it's worth in most scenarios you'll do better the more equities you have up to maybe an 80/20 split when portfolio survivability starts to be impacted by sequence or returns risk early in drawdown. If your time horizon is more than 30 years then high equity percentages really start to pull ahead. So much of retirement asset allocation comes down to the feelings of the retiree rather than the results of the analysis which is why we see cash buffers resulting from a natural fear of volatility. 

You could look at it that way, but that's not the way I designed it or use it.  For me, there is a very clear separation between my fixed income (and associated bridging funds), my drawdown portfolio and my cash as they have different purposes and timescales.  The drawdown portfolio is intended to have a time horizon of more like 70 years than 30 as I want to pass it on to my kids.  It is therefore close to 100% equities.  No portfolio survivability issues because of the fixed percentage approach, just a matter of keeping drawing percentage at or below average real returns over time to preserve the value.   I'm lucky that I can afford to have the income from it be volatile, but a cash buffer to dilute / delay the volatility is still appreciated.

Other people will have different situations and different priorities, but for those lucky enough to have more than their basic needs, I do think a multiple pot approach can be a lot easier to get your head around than trying to think of it as a single asset allocation and planning for how to change that over time.

I understand the desire to dampen volatility with cash. I'm just saying that for most people reserving cash to dampen volatility leads to a greater failure rate ie. there's more chance you'll run out of money by giving up the market gains.

This isn't supported by my own historic simulations, or by online simulators, such as https://www.2020financial.co.uk/pension-drawdown-calculator/#calculator. The failure rate actually goes down a bit rather than up.......the average end balance is lower, granted, but that's a secondary concern, and is perhaps the price for that admittedly marginally lower failure rate. Also, many retirement plans front load early withdrawals pre SP/DB.....right at the time the portfolio is most vulnerable to a poor SOR.

Of course, there are many ways to use cash in a portfolio, and so it's rather difficult to simulate them all......

US studies have shown that historically cash buffers do not improve survivability of a pension pot. It might be different for the UK.

https://bpb-us-w2.wpmucdn.com/sites.udel.edu/dist/a/855/files/2020/08/Sustainable-Withdrawal-Rates.pdf

Anyway, it's a matter of degree as everyone needs cash to spend and I do much the same as I've always done and keep a couple of years spending on hand for emergencies and liquidity.

As a larger point we are talking about managing DC pensions for income here and it's not obvious that people will have enough to support their income needs or the skills to sensibly manage their pension pot. In the US the DC pension model has been around for a couple of generations now and half of US retirees do not have enough to retire, even with the more generous US equivalent of the UK's State Pension. Many people over 65 just keep working, so maybe there needs to be more of a focus on how to use a pension pot for older people who are still working. The self directed DC pension model has failed for many US retirees because of a combination of poor implementation by government, employers and employees. The certain winners have been the financial industry.

Sorry, I have cheated and actually read the article. It uses an Indirect buffer approach and compares the % failures with simply drawing down directly from a single growth portfolio.

The buffer approach is to set up a buffer with sufficient money to pay for n years inflation linked expenditure and replenished each year from growth only if the latter has increased in value. If the buffer is exhausted expenditure is taken directly from the growth investments.

Success is determined by whether the entire portfolio survives for the specified duration of the test period.

I think this buffer approach is much like Prime Harvesting except that in the latter the bond allocation is essentially used as the buffer. Apologies if I have got this wrong.

There are many other ways of running an indirect withdrawal approach. The paper only considers one, so I cannot see it as evidence of any blanket statement about buffering.

The main difference with PH is it's a much longer term "buffer" strategy, if you want to call it that. Even in the worst bear markets it'd likely be over 10 years before you need to sell equities. The term I think is key - short term market timing is a mug's game, but longer term the market is more predictable - after all anyone invested in equities is long term timing the market - ie the assumption that over the long term equities will beat bonds/cash.

It also has researched sell tipping points, ie you don't just sell equities when they've gone up but when they're up 20% real terms. And unlike short term buffer strategies like the one linked above, it appears to generally deliver better results than a fixed asset allocation strategy. More here: Dynamic asset allocation and withdrawal in retirement - Monevator

I've not seen any evidence a short term cash buffer strategy is of any benefit (other than a comfort blanket), obviously the above article doesn't prove conclusively that none exist as the parameters could be tweaked, it's hard to prove a negative (like there's no teapot orbiting the Sun), but if evidence is that "something similar" doesn't work then surely that should be the working assumption unless something comes along to prove otherwise. 

A comfort blanket may well be useful.  In a major equity crash, or even a medium duration down turn what do you do? Cut your drawdown? Or do you carry on drawing down at the same rate regardless, confident from your use of SWR that there is no real risk?  A comfort blanket reduces the need for you to make a difficult decision for a while by which time it may not be necessary.

Personally if I was using a strategy that has been well researched and backtested in different historic market conditions including major equity crashes and prolonged bear markets, a crash on the scale of anything we've seen in the past wouldn't phase me. I'd keep calm and carry on. Why research a strategy and then panic when something happens which the strategy has been proved to cope with.

But if I was using a comfort blanket with no proven real protective effect I might panic and turn the heating down and eat dog food. 

Obviously there's always a risk that the future will be completely different to the past. Who knows, there might be a 30 year bear market a bit like in Japan from 1990, if that were to happen globally no investment portfolio containing significant equities would do well. But some would fare better than others. 

As you know, the SWR strategy has been backtested by using historical data. The initial studies with US data, and subsequent studies with international data (e.g., https://www.advisorperspectives.com/articles/2014/03/04/does-international-diversification-improve-safe-withdrawal-rates )

From that link we know that for a 30 year retirement and a 50/50 allocation to stocks and bonds, the US had a SWR of 3.7%, with New Zealand and Canada having similar values, others had slightly lower values (e.g., Sweden=3.2%, UK=3.1%, Australia=2.9%) and some (e.g., France, Italy, Germany, Austria and Japan) had a SWR below 1% (all countries significantly affected by WWII).

However, the SWR is unknown and unknowable at the start of a retirement. While history is a guide, it is not necessarily going to apply to the future. Taking the Japan example you cite, the following graph shows the MSWR for a 50/50 portfolio for a 30 year retirement in the years since 1950 as a function of start year (returns and inflation from macrohistory.net).




For a retiree in 1989, the historical record would have shown an SWR of just under 5% (the 1962 retirement would have been incomplete, but analysis in 1989 might have suggested the potential for a poor one). The retirements in the early 1970s were only 15 or so years gone so the SWR would have been uncertain in 1989. So, our retiree confidently, but conservatively, uses 4.5% as their withdrawal rate. As it then turns out, the SWR was around 4% and they ran out of money while waiting for the market to recover (see following figure which shows portfolio value in the top panel and income in the bottom panel for a 1989 retiree).

Yup. Those of course are based on a fixed asset allocation. But whatever your strategy, you can't defy gravity, if you're invested in equities and equities do really badly over a very long period then no strategy will do well. Except one that avoided equities completely.  But some do better than others. Some people think all they need to do is cope with a short term crash/recovery a bit like the COVID one. Others think you need to cope with a 10 year bear market.   

But the reality is there is no SWR. As you point out historical backtesting just tells you what would have worked in the past, not what will work in the future. These days if you want to cater for worst case scenarios, you may as well simply buy an index linked annuity as the rates are fairly similar or even better than historically tested worst case SWRs with any strategy. 

Linton

MK62 said:

Linton said:

zagfles said:

Linton said:

Bostonerimus1 said:

MK62 said:

Bostonerimus1 said:

Triumph13 said:

Bostonerimus1 said:

Triumph13 said:

zagfles said:

Triumph13 said:

I'm a big fan of a cash buffer myself, but not when used for market timing.  When coupled with a fixed withdrawal percentage, I believe (though I admit I haven't run the simulations) that it should have a lot of power to lessen the pain of spending reductions in downturns.  My plan has always been to use it to mitigate 50% of any income reduction - so if markets fall 50% my income taken only falls 25%.  A one year cash buffer would get you through four years of such a 50% drop, which seems like a reasonable reward for the cost of holding it.

More generally though, I suspect most retirees would be best to buy annuities to cover most or all of their baseline spending, making them much better placed to withstand any fluctuations in income from drawdown portfolios.  That would mean that, rather than going for a single 'default portfolio', lots of people should actually be looking to split their funds between two or more different portfolios - a traditional 'lifestyling' one for annuity purchase and an ongoing one for drawdown.  Plus a 'lifestyle to cash' one if they are planning to use the TFLS to pay off their mortgage or bridge a short gap to SP.  You really need to have gone through that thought process before you have any chance of making a sensible decision on which 'default' option to pick, and I worry about the ability of people to get through it without guidance.

If you use a cash buffer to avoid selling equities when the market is "low", then that is market timing.

Or are you saying your withdrawals are a percentage of the current pot value rather a percentage of the initial pot value? That is a very strange way of doing drawdown. If markets rose 50% would you draw 50% more? 

That's exactly what I mean, sell a fixed percentage of the portfolio.  It eliminates SORR, in return for accepting higher volatility on spending.  No market timing involved. The cash buffer then goes some way to smooth the volatility.

If equities were up 50%, I would indeed draw 50% more than planned (or at least convert equities to cash within the wrapper) , all things being equal, on the grounds that if equities were fairly valued I'd have more than enough, if they were overvalued then I'd be better with cash.  I'd then use the excess to increase the cash buffer and potentially spread it over multiple years of increased spending. 

Full disclosure, one of the reasons I can accept such high volatility is that SPs and some little DBs will eventually provide about half our income and are covered by cash in the meantime.

We are really talking about asset allocation with MMF really just being ultrashort bonds and maybe cash being zero duration fixed income. For what it's worth in most scenarios you'll do better the more equities you have up to maybe an 80/20 split when portfolio survivability starts to be impacted by sequence or returns risk early in drawdown. If your time horizon is more than 30 years then high equity percentages really start to pull ahead. So much of retirement asset allocation comes down to the feelings of the retiree rather than the results of the analysis which is why we see cash buffers resulting from a natural fear of volatility. 

You could look at it that way, but that's not the way I designed it or use it.  For me, there is a very clear separation between my fixed income (and associated bridging funds), my drawdown portfolio and my cash as they have different purposes and timescales.  The drawdown portfolio is intended to have a time horizon of more like 70 years than 30 as I want to pass it on to my kids.  It is therefore close to 100% equities.  No portfolio survivability issues because of the fixed percentage approach, just a matter of keeping drawing percentage at or below average real returns over time to preserve the value.   I'm lucky that I can afford to have the income from it be volatile, but a cash buffer to dilute / delay the volatility is still appreciated.

Other people will have different situations and different priorities, but for those lucky enough to have more than their basic needs, I do think a multiple pot approach can be a lot easier to get your head around than trying to think of it as a single asset allocation and planning for how to change that over time.

I understand the desire to dampen volatility with cash. I'm just saying that for most people reserving cash to dampen volatility leads to a greater failure rate ie. there's more chance you'll run out of money by giving up the market gains.

This isn't supported by my own historic simulations, or by online simulators, such as https://www.2020financial.co.uk/pension-drawdown-calculator/#calculator. The failure rate actually goes down a bit rather than up.......the average end balance is lower, granted, but that's a secondary concern, and is perhaps the price for that admittedly marginally lower failure rate. Also, many retirement plans front load early withdrawals pre SP/DB.....right at the time the portfolio is most vulnerable to a poor SOR.

Of course, there are many ways to use cash in a portfolio, and so it's rather difficult to simulate them all......

US studies have shown that historically cash buffers do not improve survivability of a pension pot. It might be different for the UK.

https://bpb-us-w2.wpmucdn.com/sites.udel.edu/dist/a/855/files/2020/08/Sustainable-Withdrawal-Rates.pdf

Anyway, it's a matter of degree as everyone needs cash to spend and I do much the same as I've always done and keep a couple of years spending on hand for emergencies and liquidity.

As a larger point we are talking about managing DC pensions for income here and it's not obvious that people will have enough to support their income needs or the skills to sensibly manage their pension pot. In the US the DC pension model has been around for a couple of generations now and half of US retirees do not have enough to retire, even with the more generous US equivalent of the UK's State Pension. Many people over 65 just keep working, so maybe there needs to be more of a focus on how to use a pension pot for older people who are still working. The self directed DC pension model has failed for many US retirees because of a combination of poor implementation by government, employers and employees. The certain winners have been the financial industry.

Sorry, I have cheated and actually read the article. It uses an Indirect buffer approach and compares the % failures with simply drawing down directly from a single growth portfolio.

The buffer approach is to set up a buffer with sufficient money to pay for n years inflation linked expenditure and replenished each year from growth only if the latter has increased in value. If the buffer is exhausted expenditure is taken directly from the growth investments.

Success is determined by whether the entire portfolio survives for the specified duration of the test period.

I think this buffer approach is much like Prime Harvesting except that in the latter the bond allocation is essentially used as the buffer. Apologies if I have got this wrong.

There are many other ways of running an indirect withdrawal approach. The paper only considers one, so I cannot see it as evidence of any blanket statement about buffering.

The main difference with PH is it's a much longer term "buffer" strategy, if you want to call it that. Even in the worst bear markets it'd likely be over 10 years before you need to sell equities. The term I think is key - short term market timing is a mug's game, but longer term the market is more predictable - after all anyone invested in equities is long term timing the market - ie the assumption that over the long term equities will beat bonds/cash.

It also has researched sell tipping points, ie you don't just sell equities when they've gone up but when they're up 20% real terms. And unlike short term buffer strategies like the one linked above, it appears to generally deliver better results than a fixed asset allocation strategy. More here: Dynamic asset allocation and withdrawal in retirement - Monevator

I've not seen any evidence a short term cash buffer strategy is of any benefit (other than a comfort blanket), obviously the above article doesn't prove conclusively that none exist as the parameters could be tweaked, it's hard to prove a negative (like there's no teapot orbiting the Sun), but if evidence is that "something similar" doesn't work then surely that should be the working assumption unless something comes along to prove otherwise. 

A comfort blanket may well be useful.  In a major equity crash, or even a medium duration down turn what do you do? Cut your drawdown? Or do you carry on drawing down at the same rate regardless, confident from your use of SWR that there is no real risk?  A comfort blanket reduces the need for you to make a difficult decision for a while by which time it may not be necessary.

An indirect strategy need not be short term.  I agree that short term buffer strategies are less than ideal since they are still dependent on at some point on making frequent ongoing short term transfers out of equity.

The objective driving my strategy is to keep the equity investments servicing the long term and shield them from meeting short term needs for drawdown, and to  shield the ability to spend one's money how and when one wishes from short term changes in investment returns.  To that end, like PH, one avoids taking income directly from growth investment but, unlike PH and the short term buffer approaches it also provides an integrated approach to handling one-offs by holding a large amount of stable investments. I am using the term  "indirect" to indicate it's a strategy far broader than a simple cash buffer and to avoid criticism from those who leap to wrong conclusions on hearing the word "buffer.

Removing the need for the equity tranche to provide for short term cash needs enables one to use 100% equity for long term growth and simply ride the volatility.  Like PH it does have the benefit of getting the 40% overall non-equity to do something useful.  However PH relies on using a random selection of bonds.  Are they really the best tools for the job? Is 60/40 allocation really the most appropriate.

At some point you're going to have to sell parts of that equity tranche to replenish your short/medium term pots......it can't stay in the "long term growth" pot forever......

The short term pot is replenished from a highly diversified  portfolio of income funds, both fixed interest and equity.  The income funds are bought from the gains generated by the growth funds, thus ensuring the income rises at least in line with inflation over the medium term.This means that long term equity is only sold to buy different equity thus avoiding the dangers of crystallising losses by selling for cash.

The net effect that overall the investments are highly diversified.  The key difference between this approach and most others recommended here is that each asset type is only used for the purposes to which it is best suited.  Cash is used for short term income, broad equity for long term growth, income funds for ongoing income, high interest bonds for income,  etc

OldScientist

zagfles said:

OldScientist said:

Yup. Those of course are based on a fixed asset allocation. But whatever your strategy, you can't defy gravity, if you're invested in equities and equities do really badly over a very long period then no strategy will do well. Except one that avoided equities completely.  But some do better than others. Some people think all they need to do is cope with a short term crash/recovery a bit like the COVID one. Others think you need to cope with a 10 year bear market.   

But the reality is there is no SWR. As you point out historical backtesting just tells you what would have worked in the past, not what will work in the future. These days if you want to cater for worst case scenarios, you may as well simply buy an index linked annuity as the rates are fairly similar or even better than historically tested worst case SWRs with any strategy. 

You're correct that you have to take what the market throws at you. The point is that different strategies have different failure mechanisms - SWR will provide constant income until it doesn't, when the income will fall to zero, percentage of portfolio methods will fail by providing less, and variable, income when the market is poor but the income will not fall to zero - e.g. in the Japan example, for a constant percentage of portfolio approach, the income dropped from 4.5% to 2% after 25 years before recovering by 2018. Retirees have to decide which failure mechanism is preferable and to some extent this depends on whether their essential expenditure is covered by other sources (state pension, DB pension, or, as you say, an RPI linked annuity).

westv

At the end of the day, when you retire you have a pot of money to spend.
As you can't know the future all you can be guided by is the past. A general 3% to 3.5% or so withdrawal with regular reviews is probably a good way to go. Or you could read multiple studies into why or why not you might fail but that doesn't pay the bills.

gm0

Triumph13 said:

MK62 said:

Linton said:

zagfles said:

Linton said:

Bostonerimus1 said:

MK62 said:

Bostonerimus1 said:

Triumph13 said:

Bostonerimus1 said:

Triumph13 said:

zagfles said:

Triumph13 said:

I'm a big fan of a cash buffer myself, but not when used for market timing.  When coupled with a fixed withdrawal percentage, I believe (though I admit I haven't run the simulations) that it should have a lot of power to lessen the pain of spending reductions in downturns.  My plan has always been to use it to mitigate 50% of any income reduction - so if markets fall 50% my income taken only falls 25%.  A one year cash buffer would get you through four years of such a 50% drop, which seems like a reasonable reward for the cost of holding it.

More generally though, I suspect most retirees would be best to buy annuities to cover most or all of their baseline spending, making them much better placed to withstand any fluctuations in income from drawdown portfolios.  That would mean that, rather than going for a single 'default portfolio', lots of people should actually be looking to split their funds between two or more different portfolios - a traditional 'lifestyling' one for annuity purchase and an ongoing one for drawdown.  Plus a 'lifestyle to cash' one if they are planning to use the TFLS to pay off their mortgage or bridge a short gap to SP.  You really need to have gone through that thought process before you have any chance of making a sensible decision on which 'default' option to pick, and I worry about the ability of people to get through it without guidance.

If you use a cash buffer to avoid selling equities when the market is "low", then that is market timing.

Or are you saying your withdrawals are a percentage of the current pot value rather a percentage of the initial pot value? That is a very strange way of doing drawdown. If markets rose 50% would you draw 50% more? 

That's exactly what I mean, sell a fixed percentage of the portfolio.  It eliminates SORR, in return for accepting higher volatility on spending.  No market timing involved. The cash buffer then goes some way to smooth the volatility.

If equities were up 50%, I would indeed draw 50% more than planned (or at least convert equities to cash within the wrapper) , all things being equal, on the grounds that if equities were fairly valued I'd have more than enough, if they were overvalued then I'd be better with cash.  I'd then use the excess to increase the cash buffer and potentially spread it over multiple years of increased spending. 

Full disclosure, one of the reasons I can accept such high volatility is that SPs and some little DBs will eventually provide about half our income and are covered by cash in the meantime.

We are really talking about asset allocation with MMF really just being ultrashort bonds and maybe cash being zero duration fixed income. For what it's worth in most scenarios you'll do better the more equities you have up to maybe an 80/20 split when portfolio survivability starts to be impacted by sequence or returns risk early in drawdown. If your time horizon is more than 30 years then high equity percentages really start to pull ahead. So much of retirement asset allocation comes down to the feelings of the retiree rather than the results of the analysis which is why we see cash buffers resulting from a natural fear of volatility. 

You could look at it that way, but that's not the way I designed it or use it.  For me, there is a very clear separation between my fixed income (and associated bridging funds), my drawdown portfolio and my cash as they have different purposes and timescales.  The drawdown portfolio is intended to have a time horizon of more like 70 years than 30 as I want to pass it on to my kids.  It is therefore close to 100% equities.  No portfolio survivability issues because of the fixed percentage approach, just a matter of keeping drawing percentage at or below average real returns over time to preserve the value.   I'm lucky that I can afford to have the income from it be volatile, but a cash buffer to dilute / delay the volatility is still appreciated.

Other people will have different situations and different priorities, but for those lucky enough to have more than their basic needs, I do think a multiple pot approach can be a lot easier to get your head around than trying to think of it as a single asset allocation and planning for how to change that over time.

I understand the desire to dampen volatility with cash. I'm just saying that for most people reserving cash to dampen volatility leads to a greater failure rate ie. there's more chance you'll run out of money by giving up the market gains.

This isn't supported by my own historic simulations, or by online simulators, such as https://www.2020financial.co.uk/pension-drawdown-calculator/#calculator. The failure rate actually goes down a bit rather than up.......the average end balance is lower, granted, but that's a secondary concern, and is perhaps the price for that admittedly marginally lower failure rate. Also, many retirement plans front load early withdrawals pre SP/DB.....right at the time the portfolio is most vulnerable to a poor SOR.

Of course, there are many ways to use cash in a portfolio, and so it's rather difficult to simulate them all......

US studies have shown that historically cash buffers do not improve survivability of a pension pot. It might be different for the UK.

https://bpb-us-w2.wpmucdn.com/sites.udel.edu/dist/a/855/files/2020/08/Sustainable-Withdrawal-Rates.pdf

Anyway, it's a matter of degree as everyone needs cash to spend and I do much the same as I've always done and keep a couple of years spending on hand for emergencies and liquidity.

As a larger point we are talking about managing DC pensions for income here and it's not obvious that people will have enough to support their income needs or the skills to sensibly manage their pension pot. In the US the DC pension model has been around for a couple of generations now and half of US retirees do not have enough to retire, even with the more generous US equivalent of the UK's State Pension. Many people over 65 just keep working, so maybe there needs to be more of a focus on how to use a pension pot for older people who are still working. The self directed DC pension model has failed for many US retirees because of a combination of poor implementation by government, employers and employees. The certain winners have been the financial industry.

Sorry, I have cheated and actually read the article. It uses an Indirect buffer approach and compares the % failures with simply drawing down directly from a single growth portfolio.

The buffer approach is to set up a buffer with sufficient money to pay for n years inflation linked expenditure and replenished each year from growth only if the latter has increased in value. If the buffer is exhausted expenditure is taken directly from the growth investments.

Success is determined by whether the entire portfolio survives for the specified duration of the test period.

I think this buffer approach is much like Prime Harvesting except that in the latter the bond allocation is essentially used as the buffer. Apologies if I have got this wrong.

There are many other ways of running an indirect withdrawal approach. The paper only considers one, so I cannot see it as evidence of any blanket statement about buffering.

The main difference with PH is it's a much longer term "buffer" strategy, if you want to call it that. Even in the worst bear markets it'd likely be over 10 years before you need to sell equities. The term I think is key - short term market timing is a mug's game, but longer term the market is more predictable - after all anyone invested in equities is long term timing the market - ie the assumption that over the long term equities will beat bonds/cash.

It also has researched sell tipping points, ie you don't just sell equities when they've gone up but when they're up 20% real terms. And unlike short term buffer strategies like the one linked above, it appears to generally deliver better results than a fixed asset allocation strategy. More here: Dynamic asset allocation and withdrawal in retirement - Monevator

I've not seen any evidence a short term cash buffer strategy is of any benefit (other than a comfort blanket), obviously the above article doesn't prove conclusively that none exist as the parameters could be tweaked, it's hard to prove a negative (like there's no teapot orbiting the Sun), but if evidence is that "something similar" doesn't work then surely that should be the working assumption unless something comes along to prove otherwise. 

A comfort blanket may well be useful.  In a major equity crash, or even a medium duration down turn what do you do? Cut your drawdown? Or do you carry on drawing down at the same rate regardless, confident from your use of SWR that there is no real risk?  A comfort blanket reduces the need for you to make a difficult decision for a while by which time it may not be necessary.

An indirect strategy need not be short term.  I agree that short term buffer strategies are less than ideal since they are still dependent on at some point on making frequent ongoing short term transfers out of equity.

The objective driving my strategy is to keep the equity investments servicing the long term and shield them from meeting short term needs for drawdown, and to  shield the ability to spend one's money how and when one wishes from short term changes in investment returns.  To that end, like PH, one avoids taking income directly from growth investment but, unlike PH and the short term buffer approaches it also provides an integrated approach to handling one-offs by holding a large amount of stable investments. I am using the term  "indirect" to indicate it's a strategy far broader than a simple cash buffer and to avoid criticism from those who leap to wrong conclusions on hearing the word "buffer.

Removing the need for the equity tranche to provide for short term cash needs enables one to use 100% equity for long term growth and simply ride the volatility.  Like PH it does have the benefit of getting the 40% overall non-equity to do something useful.  However PH relies on using a random selection of bonds.  Are they really the best tools for the job? Is 60/40 allocation really the most appropriate.

At some point you're going to have to sell parts of that equity tranche to replenish your short/medium term pots......it can't stay in the "long term growth" pot forever......

Which is another reason I favour selling a fixed percentage of equities every year.  I benefitted from dollar cost averaging on the way in, now I benefit from something similar on the wayout too

Errr. not really - this idea depends - a lot - on the comparator you choose to feel good about it. 

A good way to characterise dollar cost averaging positively is during accumulation.  Which is clearly a fine thing for saving up a pension month by month over decades.  Dips are your friend.  But it is the transfer of wealth from the selller of dips to the buyer of them.  The benefit for the buyer which most acknowledge is real -  of buying through volatility is not magic.  It comes from somewhere.  And the somewhere it comes from is the selling of cheaper units of the same funds during said volatility by the (forced) seller for income or any other deaccumulator of growth assets of that class.

Clearly selling a large chunk in a dip could well be a panic response and signficantly worse than just selling regularly month by month.  So if you make *that* the comparator.  Monthly selling drip drip drip looks fine.

Nonetheless - my own preference is for selling funds only at rebalancing i.e. once in 12-18 months.  So the impact of short term volatility in terms of selling prices achieved between rebalancings is precisely zero.  I don't mind having some cash for income - as I intend to hold some sequence risk in cash/short gilts/MMF anyway.  So income will always be coming from "not equities" sources.

This is superior *for my purposes* - to the platform auto selling equities to cash month by month on a fixed day and getting market best execution.   Flash crash - oh dear.   Volatility dip like 2020 - oh dear.  

So income comes from buffer.  Controlled rebalancing timing, manual trades.  Equity sales happen where they happen at all at my choice of timing.  And using limit orders.

You choose the set and forget mechanism you like - if you are a set and forget drawdown investor.  Or you manage things more actively.  There is no "best" as such - only best for your goals and preferences and the risks you attempt to mitigate and those you carry

Bostonerimus1

OldScientist said:

MK62 said:

Bostonerimus1 said:

Triumph13 said:

Bostonerimus1 said:

Triumph13 said:

zagfles said:

Triumph13 said:

I'm a big fan of a cash buffer myself, but not when used for market timing.  When coupled with a fixed withdrawal percentage, I believe (though I admit I haven't run the simulations) that it should have a lot of power to lessen the pain of spending reductions in downturns.  My plan has always been to use it to mitigate 50% of any income reduction - so if markets fall 50% my income taken only falls 25%.  A one year cash buffer would get you through four years of such a 50% drop, which seems like a reasonable reward for the cost of holding it.

More generally though, I suspect most retirees would be best to buy annuities to cover most or all of their baseline spending, making them much better placed to withstand any fluctuations in income from drawdown portfolios.  That would mean that, rather than going for a single 'default portfolio', lots of people should actually be looking to split their funds between two or more different portfolios - a traditional 'lifestyling' one for annuity purchase and an ongoing one for drawdown.  Plus a 'lifestyle to cash' one if they are planning to use the TFLS to pay off their mortgage or bridge a short gap to SP.  You really need to have gone through that thought process before you have any chance of making a sensible decision on which 'default' option to pick, and I worry about the ability of people to get through it without guidance.

If you use a cash buffer to avoid selling equities when the market is "low", then that is market timing.

Or are you saying your withdrawals are a percentage of the current pot value rather a percentage of the initial pot value? That is a very strange way of doing drawdown. If markets rose 50% would you draw 50% more? 

That's exactly what I mean, sell a fixed percentage of the portfolio.  It eliminates SORR, in return for accepting higher volatility on spending.  No market timing involved. The cash buffer then goes some way to smooth the volatility.

If equities were up 50%, I would indeed draw 50% more than planned (or at least convert equities to cash within the wrapper) , all things being equal, on the grounds that if equities were fairly valued I'd have more than enough, if they were overvalued then I'd be better with cash.  I'd then use the excess to increase the cash buffer and potentially spread it over multiple years of increased spending. 

Full disclosure, one of the reasons I can accept such high volatility is that SPs and some little DBs will eventually provide about half our income and are covered by cash in the meantime.

We are really talking about asset allocation with MMF really just being ultrashort bonds and maybe cash being zero duration fixed income. For what it's worth in most scenarios you'll do better the more equities you have up to maybe an 80/20 split when portfolio survivability starts to be impacted by sequence or returns risk early in drawdown. If your time horizon is more than 30 years then high equity percentages really start to pull ahead. So much of retirement asset allocation comes down to the feelings of the retiree rather than the results of the analysis which is why we see cash buffers resulting from a natural fear of volatility. 

You could look at it that way, but that's not the way I designed it or use it.  For me, there is a very clear separation between my fixed income (and associated bridging funds), my drawdown portfolio and my cash as they have different purposes and timescales.  The drawdown portfolio is intended to have a time horizon of more like 70 years than 30 as I want to pass it on to my kids.  It is therefore close to 100% equities.  No portfolio survivability issues because of the fixed percentage approach, just a matter of keeping drawing percentage at or below average real returns over time to preserve the value.   I'm lucky that I can afford to have the income from it be volatile, but a cash buffer to dilute / delay the volatility is still appreciated.

Other people will have different situations and different priorities, but for those lucky enough to have more than their basic needs, I do think a multiple pot approach can be a lot easier to get your head around than trying to think of it as a single asset allocation and planning for how to change that over time.

I understand the desire to dampen volatility with cash. I'm just saying that for most people reserving cash to dampen volatility leads to a greater failure rate ie. there's more chance you'll run out of money by giving up the market gains.

This isn't supported by my own historic simulations, or by online simulators, such as https://www.2020financial.co.uk/pension-drawdown-calculator/#calculator. The failure rate actually goes down a bit rather than up.......the average end balance is lower, granted, but that's a secondary concern, and is perhaps the price for that admittedly marginally lower failure rate. Also, many retirement plans front load early withdrawals pre SP/DB.....right at the time the portfolio is most vulnerable to a poor SOR.

Of course, there are many ways to use cash in a portfolio, and so it's rather difficult to simulate them all......

One point about the 2020financial calculator is that it using the return data from the Barclay's equity gilt study. The bonds in that study are undated (e.g., Consols) before about 1960, 20 year maturities until 1980 or so, and 15 year maturity after that. In other words, the bonds being simulated are far longer maturity than usually held during retirement.

The effect of gilt maturity on SWR in the UK can be found at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4742456 and for percentage of portfolio withdrawals at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4827947

For full disclosure, I am the author of those two papers.

Nice! Ive always held intermediate term bond funds as a "middle course" between interest rate variations. Well I've always thought of them as intermediate average maturity, but you might differ. I used Vanguard Total Bond Index for a long time and that has an average maturity of 8.5 years and duration of 6 years. I don't own it now but do own the multi-asset Vanguard Wellesley fund and it's bonds have an average maturity of 10.4 years and duration 6.7 years. It's an income fund so they go a little longer. I also own the Vanguard Federal MMF which has an average maturity of 11 days.

Triumph13

gm0 said:

Triumph13 said:

MK62 said:

Linton said:

zagfles said:

Linton said:

Bostonerimus1 said:

MK62 said:

Bostonerimus1 said:

Triumph13 said:

Bostonerimus1 said:

Triumph13 said:

zagfles said:

Triumph13 said:

I'm a big fan of a cash buffer myself, but not when used for market timing.  When coupled with a fixed withdrawal percentage, I believe (though I admit I haven't run the simulations) that it should have a lot of power to lessen the pain of spending reductions in downturns.  My plan has always been to use it to mitigate 50% of any income reduction - so if markets fall 50% my income taken only falls 25%.  A one year cash buffer would get you through four years of such a 50% drop, which seems like a reasonable reward for the cost of holding it.

More generally though, I suspect most retirees would be best to buy annuities to cover most or all of their baseline spending, making them much better placed to withstand any fluctuations in income from drawdown portfolios.  That would mean that, rather than going for a single 'default portfolio', lots of people should actually be looking to split their funds between two or more different portfolios - a traditional 'lifestyling' one for annuity purchase and an ongoing one for drawdown.  Plus a 'lifestyle to cash' one if they are planning to use the TFLS to pay off their mortgage or bridge a short gap to SP.  You really need to have gone through that thought process before you have any chance of making a sensible decision on which 'default' option to pick, and I worry about the ability of people to get through it without guidance.

If you use a cash buffer to avoid selling equities when the market is "low", then that is market timing.

Or are you saying your withdrawals are a percentage of the current pot value rather a percentage of the initial pot value? That is a very strange way of doing drawdown. If markets rose 50% would you draw 50% more? 

That's exactly what I mean, sell a fixed percentage of the portfolio.  It eliminates SORR, in return for accepting higher volatility on spending.  No market timing involved. The cash buffer then goes some way to smooth the volatility.

If equities were up 50%, I would indeed draw 50% more than planned (or at least convert equities to cash within the wrapper) , all things being equal, on the grounds that if equities were fairly valued I'd have more than enough, if they were overvalued then I'd be better with cash.  I'd then use the excess to increase the cash buffer and potentially spread it over multiple years of increased spending. 

Full disclosure, one of the reasons I can accept such high volatility is that SPs and some little DBs will eventually provide about half our income and are covered by cash in the meantime.

We are really talking about asset allocation with MMF really just being ultrashort bonds and maybe cash being zero duration fixed income. For what it's worth in most scenarios you'll do better the more equities you have up to maybe an 80/20 split when portfolio survivability starts to be impacted by sequence or returns risk early in drawdown. If your time horizon is more than 30 years then high equity percentages really start to pull ahead. So much of retirement asset allocation comes down to the feelings of the retiree rather than the results of the analysis which is why we see cash buffers resulting from a natural fear of volatility. 

You could look at it that way, but that's not the way I designed it or use it.  For me, there is a very clear separation between my fixed income (and associated bridging funds), my drawdown portfolio and my cash as they have different purposes and timescales.  The drawdown portfolio is intended to have a time horizon of more like 70 years than 30 as I want to pass it on to my kids.  It is therefore close to 100% equities.  No portfolio survivability issues because of the fixed percentage approach, just a matter of keeping drawing percentage at or below average real returns over time to preserve the value.   I'm lucky that I can afford to have the income from it be volatile, but a cash buffer to dilute / delay the volatility is still appreciated.

Other people will have different situations and different priorities, but for those lucky enough to have more than their basic needs, I do think a multiple pot approach can be a lot easier to get your head around than trying to think of it as a single asset allocation and planning for how to change that over time.

I understand the desire to dampen volatility with cash. I'm just saying that for most people reserving cash to dampen volatility leads to a greater failure rate ie. there's more chance you'll run out of money by giving up the market gains.

This isn't supported by my own historic simulations, or by online simulators, such as https://www.2020financial.co.uk/pension-drawdown-calculator/#calculator. The failure rate actually goes down a bit rather than up.......the average end balance is lower, granted, but that's a secondary concern, and is perhaps the price for that admittedly marginally lower failure rate. Also, many retirement plans front load early withdrawals pre SP/DB.....right at the time the portfolio is most vulnerable to a poor SOR.

Of course, there are many ways to use cash in a portfolio, and so it's rather difficult to simulate them all......

US studies have shown that historically cash buffers do not improve survivability of a pension pot. It might be different for the UK.

https://bpb-us-w2.wpmucdn.com/sites.udel.edu/dist/a/855/files/2020/08/Sustainable-Withdrawal-Rates.pdf

Anyway, it's a matter of degree as everyone needs cash to spend and I do much the same as I've always done and keep a couple of years spending on hand for emergencies and liquidity.

As a larger point we are talking about managing DC pensions for income here and it's not obvious that people will have enough to support their income needs or the skills to sensibly manage their pension pot. In the US the DC pension model has been around for a couple of generations now and half of US retirees do not have enough to retire, even with the more generous US equivalent of the UK's State Pension. Many people over 65 just keep working, so maybe there needs to be more of a focus on how to use a pension pot for older people who are still working. The self directed DC pension model has failed for many US retirees because of a combination of poor implementation by government, employers and employees. The certain winners have been the financial industry.

Sorry, I have cheated and actually read the article. It uses an Indirect buffer approach and compares the % failures with simply drawing down directly from a single growth portfolio.

The buffer approach is to set up a buffer with sufficient money to pay for n years inflation linked expenditure and replenished each year from growth only if the latter has increased in value. If the buffer is exhausted expenditure is taken directly from the growth investments.

Success is determined by whether the entire portfolio survives for the specified duration of the test period.

I think this buffer approach is much like Prime Harvesting except that in the latter the bond allocation is essentially used as the buffer. Apologies if I have got this wrong.

There are many other ways of running an indirect withdrawal approach. The paper only considers one, so I cannot see it as evidence of any blanket statement about buffering.

The main difference with PH is it's a much longer term "buffer" strategy, if you want to call it that. Even in the worst bear markets it'd likely be over 10 years before you need to sell equities. The term I think is key - short term market timing is a mug's game, but longer term the market is more predictable - after all anyone invested in equities is long term timing the market - ie the assumption that over the long term equities will beat bonds/cash.

It also has researched sell tipping points, ie you don't just sell equities when they've gone up but when they're up 20% real terms. And unlike short term buffer strategies like the one linked above, it appears to generally deliver better results than a fixed asset allocation strategy. More here: Dynamic asset allocation and withdrawal in retirement - Monevator

I've not seen any evidence a short term cash buffer strategy is of any benefit (other than a comfort blanket), obviously the above article doesn't prove conclusively that none exist as the parameters could be tweaked, it's hard to prove a negative (like there's no teapot orbiting the Sun), but if evidence is that "something similar" doesn't work then surely that should be the working assumption unless something comes along to prove otherwise. 

A comfort blanket may well be useful.  In a major equity crash, or even a medium duration down turn what do you do? Cut your drawdown? Or do you carry on drawing down at the same rate regardless, confident from your use of SWR that there is no real risk?  A comfort blanket reduces the need for you to make a difficult decision for a while by which time it may not be necessary.

An indirect strategy need not be short term.  I agree that short term buffer strategies are less than ideal since they are still dependent on at some point on making frequent ongoing short term transfers out of equity.

The objective driving my strategy is to keep the equity investments servicing the long term and shield them from meeting short term needs for drawdown, and to  shield the ability to spend one's money how and when one wishes from short term changes in investment returns.  To that end, like PH, one avoids taking income directly from growth investment but, unlike PH and the short term buffer approaches it also provides an integrated approach to handling one-offs by holding a large amount of stable investments. I am using the term  "indirect" to indicate it's a strategy far broader than a simple cash buffer and to avoid criticism from those who leap to wrong conclusions on hearing the word "buffer.

Removing the need for the equity tranche to provide for short term cash needs enables one to use 100% equity for long term growth and simply ride the volatility.  Like PH it does have the benefit of getting the 40% overall non-equity to do something useful.  However PH relies on using a random selection of bonds.  Are they really the best tools for the job? Is 60/40 allocation really the most appropriate.

At some point you're going to have to sell parts of that equity tranche to replenish your short/medium term pots......it can't stay in the "long term growth" pot forever......

Which is another reason I favour selling a fixed percentage of equities every year.  I benefitted from dollar cost averaging on the way in, now I benefit from something similar on the wayout too

Errr. not really - this idea depends - a lot - on the comparator you choose to feel good about it. 

A good way to characterise dollar cost averaging positively is during accumulation.  Which is clearly a fine thing for saving up a pension month by month over decades.  Dips are your friend.  But it is the transfer of wealth from the selller of dips to the buyer of them.  The benefit for the buyer which most acknowledge is real -  of buying through volatility is not magic.  It comes from somewhere.  And the somewhere it comes from is the selling of cheaper units of the same funds during said volatility by the (forced) seller for income or any other deaccumulator of growth assets of that class.

Clearly selling a large chunk in a dip could well be a panic response and signficantly worse than just selling regularly month by month.  So if you make *that* the comparator.  Monthly selling drip drip drip looks fine.

Nonetheless - my own preference is for selling funds only at rebalancing i.e. once in 12-18 months.  So the impact of short term volatility in terms of selling prices achieved between rebalancings is precisely zero.  I don't mind having some cash for income - as I intend to hold some sequence risk in cash/short gilts/MMF anyway.  So income will always be coming from "not equities" sources.

This is superior *for my purposes* - to the platform auto selling equities to cash month by month on a fixed day and getting market best execution.   Flash crash - oh dear.   Volatility dip like 2020 - oh dear.  

So income comes from buffer.  Controlled rebalancing timing, manual trades.  Equity sales happen where they happen at all at my choice of timing.  And using limit orders.

You choose the set and forget mechanism you like - if you are a set and forget drawdown investor.  Or you manage things more actively.  There is no "best" as such - only best for your goals and preferences and the risks you attempt to mitigate and those you carry

The comparator is to someone selling fixed dollar amounts of stocks at a regular interval - whatever that interval may be.  Because I'm selling fixed portfolio percentages, over time a larger proportion of my cash will end up coming from times when markets were high than when markets were low.  It's as simple as that.

If you think you can beat that by market timing, then so be it.

OldScientist

westv said:

At the end of the day, when you retire you have a pot of money to spend.
As you can't know the future all you can be guided by is the past. A general 3% to 3.5% or so withdrawal with regular reviews is probably a good way to go. Or you could read multiple studies into why or why not you might fail but that doesn't pay the bills.

Although, reading multiple studies and then implementing a robust withdrawal approach may actually pay the bills

I'd agree that your suggested method is one of very many sensible approaches since it is essentially a dynamic withdrawal strategy. Personally, I think it would probably be useful to establish the review frequency (or criteria) and the criteria for making changes to withdrawals beforehand to minimise behavioural risk.

mark55man

Dazza1902 said:

It might be helpful for the op to give some details, age, how long he needs to bridge to Sp, what other income he may have, what his spends will be at retirement etc.
It sounds like you are looking for a simple way to draw income. At its simplest there are monthly paying dividend focused funds , rather than having to sell assets for income. It's not popular with those on here.

So I think @Dazza1902 has a point.  The fascinating conversations about SWR GK PH etc answer the meta question about techniques to make a pot of money last (and the evidence of success in different techniques).  But I think the OP might have been after an understanding of specific income oriented funds.  Some of the options have been discussed - annuities, bond ladders etc.  Funds that seem to be missing are Income (or income and growth) oriented investment trusts, also high yield bond funds, or indeed fixed interest bonds.  I understand that high yield carries risks to total return and income volatility as yields dwindle for investors who just want an Annuity like (lite?) experience. 

In my case I have an adequate DB to cover my essentials so just want a flexible income but as high as possible especially early prior to state pension age.  Alternatively, on the lemon fool website they have decades long threads about construction and management of high yield portfolios (and equally long arguments about the merits of dividends vs total return)