SWR Question

JohnWinder

Croeso69 said:

I dont like bonds in the current environment. How good is G-K for 95% equity and 5% cash to fund drawdown?

Isn't that what firesim does beautifully? It models exactly that (if you specify it) over all the 30 year periods for which there is data.

Croeso69

BritishInvestor said:

Croeso69 said:

jamesd said:

3.7% is the UK equivalent of fifty percent large cap and fifty percent intermediate bonds in the US. But both are before costs and the adjustment for that is to deduct a third of costs. And of course it's for thirty years, not fifty, so it is a bit high.

Guyton-Klinger at 99% success for for forty years starts at 5.5% with a 65% equities and 35% bonds mix.

I usually give 3.2% and 5% to allow for 1.5% total costs.

I dont like bonds in the current environment. How good is G-K for 95% equity and 5% cash to fund drawdown?

I think you really have to understand the potential downside if considering this approach. 

At current levels there is a lot more downside in bonds i think?

michaels

Croeso69 said:

BritishInvestor said:

Croeso69 said:

jamesd said:

3.7% is the UK equivalent of fifty percent large cap and fifty percent intermediate bonds in the US. But both are before costs and the adjustment for that is to deduct a third of costs. And of course it's for thirty years, not fifty, so it is a bit high.

Guyton-Klinger at 99% success for for forty years starts at 5.5% with a 65% equities and 35% bonds mix.

I usually give 3.2% and 5% to allow for 1.5% total costs.

I dont like bonds in the current environment. How good is G-K for 95% equity and 5% cash to fund drawdown?

I think you really have to understand the potential downside if considering this approach. 

At current levels there is a lot more downside in bonds i think?

What concerns me is that in the past bond and equity returns were to an extent inversely correlated but that going forward we might see a period of positive correlation where both equities and bonds fall in value. I know 150 years of data is a long time but if you were running a simulation you would run 10s of thousands of scenarios. Also historically there has also been an element of reversion to mean for asset values but I am not sure there is any reason this is rule rather than just an observation.

jamesd

michaels said:

jamesd said:

Likely to add a BTL flat soon and many people will hate that, though it'll start out unleveraged in current equity market conditions.

Given borrowing rates are so low, why would you eschew leverage in a property investment?  Obviously the amount of 'management time' and 'individual risk' are much higher for a property investment than investment funds.

It's a house converted into two flats. I own half the freehold and the leasehold of one flat already.

Vendor bought the other one back in the 70s but hasn't used it for anything other than storage for a couple of decades, so will have quite a CGT bill. So we're considering doing the transaction over several years so I can deliver him money value in reduced CGT without spending money.

Not sure it's regularly mortgageable at the moment either. There's some water penetration under a flat roof at the rear that needs fixing. Half of the cost of that work will show up in a price discount. Vendor prefers to pass that work on to me.

Also lots of general refurbishment needed, like adding central heating and double or triple glazing.

Once the work is done and the mortgage value is higher I'll get one. Perhaps by then equities will have become more desirable.

michaels said:

Finally we are about to get thrown out of ratesetter as they have sold their book - which platforms are you using if you don't mind sharing - are any of them sippable?

Ablrate and unbolted are the two I use most. Ablrate is profitable. Unbolted well funded but not yet profitable with a big legal case in progress, so there's more platform risk. Don't know about SIPP.

BritishInvestor

Croeso69 said:

BritishInvestor said:

Croeso69 said:

jamesd said:

3.7% is the UK equivalent of fifty percent large cap and fifty percent intermediate bonds in the US. But both are before costs and the adjustment for that is to deduct a third of costs. And of course it's for thirty years, not fifty, so it is a bit high.

Guyton-Klinger at 99% success for for forty years starts at 5.5% with a 65% equities and 35% bonds mix.

I usually give 3.2% and 5% to allow for 1.5% total costs.

I dont like bonds in the current environment. How good is G-K for 95% equity and 5% cash to fund drawdown?

I think you really have to understand the potential downside if considering this approach. 

At current levels there is a lot more downside in bonds i think?

Downsides in what sense?

Croeso69

BritishInvestor said:

Croeso69 said:

BritishInvestor said:

Croeso69 said:

jamesd said:

3.7% is the UK equivalent of fifty percent large cap and fifty percent intermediate bonds in the US. But both are before costs and the adjustment for that is to deduct a third of costs. And of course it's for thirty years, not fifty, so it is a bit high.

Guyton-Klinger at 99% success for for forty years starts at 5.5% with a 65% equities and 35% bonds mix.

I usually give 3.2% and 5% to allow for 1.5% total costs.

I dont like bonds in the current environment. How good is G-K for 95% equity and 5% cash to fund drawdown?

I think you really have to understand the potential downside if considering this approach. 

At current levels there is a lot more downside in bonds i think?

Downsides in what sense?

Long term loss of capital.

BritishInvestor

Croeso69 said:

BritishInvestor said:

Croeso69 said:

BritishInvestor said:

Croeso69 said:

jamesd said:

3.7% is the UK equivalent of fifty percent large cap and fifty percent intermediate bonds in the US. But both are before costs and the adjustment for that is to deduct a third of costs. And of course it's for thirty years, not fifty, so it is a bit high.

Guyton-Klinger at 99% success for for forty years starts at 5.5% with a 65% equities and 35% bonds mix.

I usually give 3.2% and 5% to allow for 1.5% total costs.

I dont like bonds in the current environment. How good is G-K for 95% equity and 5% cash to fund drawdown?

I think you really have to understand the potential downside if considering this approach. 

At current levels there is a lot more downside in bonds i think?

Downsides in what sense?

Long term loss of capital.

Yep, but if you bail out because you can't taken a >50% drawdown then there won't be a long term.

Croeso69

BritishInvestor said:

Croeso69 said:

BritishInvestor said:

Croeso69 said:

BritishInvestor said:

Croeso69 said:

jamesd said:

3.7% is the UK equivalent of fifty percent large cap and fifty percent intermediate bonds in the US. But both are before costs and the adjustment for that is to deduct a third of costs. And of course it's for thirty years, not fifty, so it is a bit high.

Guyton-Klinger at 99% success for for forty years starts at 5.5% with a 65% equities and 35% bonds mix.

I usually give 3.2% and 5% to allow for 1.5% total costs.

I dont like bonds in the current environment. How good is G-K for 95% equity and 5% cash to fund drawdown?

I think you really have to understand the potential downside if considering this approach. 

At current levels there is a lot more downside in bonds i think?

Downsides in what sense?

Long term loss of capital.

Yep, but if you bail out because you can't taken a >50% drawdown then there won't be a long term.

I am not bailing. I bought Legal and General at 50p before the crash of 2007/08 which saw them fall to approx 20p. I held on and sold a few years later at a decent profit. Wish i had bought more to average down. Ho hum. Now I only use IT, ETF and OEICs.

gm0

ac198179

Hopefully that's cleared up your original query.  ;->

We don't know.  Whether the future resembles the past - but we don't have a better model that we could feasibly agree on
Backtesting is not proof of safety. But we have nothing meaningfully better to provide a guide to setting up WR for deaccumulation. (Not starting an argument about backtesting and MC sim - honest).

Of course you can make SWR "safer" using 100% - no failures for the design life 30, 40, other, in backtesting - so the value that emerges backtests for the worst possible start dates and sequences for your desired asset mix i.e. the next bit higher would be the point the first "terrible event from history" pops up and a cohort runs out of money early.

You can stick to 40-~70% equities which seems to be a prudent range (from backtesting).  Above 70 more volatility driven depletion before death failures statistically appear though this of course depends upon how you set, take, index income). Also more people die with even huger pots at 80%-100% which is little comfort if you are the other one with the crash + prolonged slump to deal with at the start of 40 year retirement and a high income need vs pot size.  Exactly the person who then sells for income during the slump at low prices and doesn't have the assets when returns recover later.  Limps home. Low variable income. Or actual early depletion in extremis.

You can make it even safer by examining a given WR plan's behaviours with tougher markets again.  Various academic studies and books do exactly this. 

But nothing is ever safe even with backtesting set to "no failures" - because historical risk is not all risk.  Taleb (Black Swans aka Fat Tails).  Future not being like the past (a possibility even if this is predicted wrongly far more often than it actually happens).  

If you REALLY want to know more about what's under the covers of the 4% rule and its UK adaptations then take a look at the ERN blog if you haven't already.  WR series.  Monevator has some good pages on it also.  And for the "need to understand the detail before I can be comfortable people" then there is always Michael McClung (Living off your Money) who does the whole methods and testing and literature survey and stress test design thing to exhaustion*.  *His and most readers.

This whole line of safety thinking lowers your WR which may be OK for you or not based on income expectation and pot size.  You can reduce the possibility of early depletion and having to worry about longevity.  But you will likely set yourself up for a large pot left over at 95 in the majority of cases.  Tuning backtesting may improve precision but not necessarily predictive accuracy.

ac198179

gm0 said:

ac198179

Hopefully that's cleared up your original query.  ;->

We don't know.  Whether the future resembles the past - but we don't have a better model that we could feasibly agree on
Backtesting is not proof of safety. But we have nothing meaningfully better to provide a guide to setting up WR for deaccumulation. (Not starting an argument about backtesting and MC sim - honest).

Of course you can make SWR "safer" using 100% - no failures for the design life 30, 40, other, in backtesting - so the value that emerges backtests for the worst possible start dates and sequences for your desired asset mix i.e. the next bit higher would be the point the first "terrible event from history" pops up and a cohort runs out of money early.

You can stick to 40-~70% equities which seems to be a prudent range (from backtesting).  Above 70 more volatility driven depletion before death failures statistically appear though this of course depends upon how you set, take, index income). Also more people die with even huger pots at 80%-100% which is little comfort if you are the other one with the crash + prolonged slump to deal with at the start of 40 year retirement and a high income need vs pot size.  Exactly the person who then sells for income during the slump at low prices and doesn't have the assets when returns recover later.  Limps home. Low variable income. Or actual early depletion in extremis.

You can make it even safer by examining a given WR plan's behaviours with tougher markets again.  Various academic studies and books do exactly this. 

But nothing is ever safe even with backtesting set to "no failures" - because historical risk is not all risk.  Taleb (Black Swans aka Fat Tails).  Future not being like the past (a possibility even if this is predicted wrongly far more often than it actually happens).  

If you REALLY want to know more about what's under the covers of the 4% rule and its UK adaptations then take a look at the ERN blog if you haven't already.  WR series.  Monevator has some good pages on it also.  And for the "need to understand the detail before I can be comfortable people" then there is always Michael McClung (Living off your Money) who does the whole methods and testing and literature survey and stress test design thing to exhaustion*.  *His and most readers.

This whole line of safety thinking lowers your WR which may be OK for you or not based on income expectation and pot size.  You can reduce the possibility of early depletion and having to worry about longevity.  But you will likely set yourself up for a large pot left over at 95 in the majority of cases.  Tuning backtesting may improve precision but not necessarily predictive accuracy.

Thanks very much & thanks again to everyone else, lots of very useful info & discussion in this thread for me to wade though/digest.