risk profiling - lower risk actually = higher risk

Pat38493

Deleted_User said:

Pat38493 said:

Linton said:

Pat38493 said:

Mistermeaner said:

hi

question regards risk profiling

I understand that a number of investment companies / advisors will work with clients to understand their 'risk profile' and then structure their invetsments accordingly

i.e. someone 'risk adverse' would be put into generally lower volatility products such as bond weighted funds where as someone with a higher risk appetite would be higher in equities etc.

while i get the principle of this is it not somewhat self defeating as over a long term (esp with pensions) regardless of your 'risk tollerance' going bond heavy at age 20, 30 even 40 and 50 is simply lowering your returns and therefore exposing you to greater risks

Isn;t the concept just misplaced?

Reason for asking my company is moving to a new DCpension provider and the new company will be offering a 'risk review' to help people decide on their investments - i see the whole process as somewhat arbitrary and flawed

Appreciate thoughts and comments 


Thanks

As you can see from perusing these boards, a lot of people tend to get a bit of a panic on if they see their pension fund making precipitous drops in value, so if your pension fund dropped by 40% when you were one year into retirement, can you still stay rational enough to assume that this is still within the parameters of past history that you tested against, and therefore you don't need to take any action?  

Even if you know your strategy is fully rational, if you are the type of person who will lose a lot of sleep when your pension has dropped a lot early in retirement, I guess this is why people need profiling - you might be wealthier in the long run but you might not be happier if you are losing lot of sleep worrying about the recent market crash.

However fundamentally you are not far off - putting too much of your funds into low valatility and low risk investments is likely to end up with you working a lot longer.  

I know a couple of people who are fully aware of all this, but they still keep bringing up how much their fund dropped this year - go figure....

Rationally I suppose you should only start worrying if the markets do things that have never happened before and you were not prepared for in your retirement planning.

I think this is putting too much faith in the parameters of history.

In practice we have no really reliable parameters of history.  Take a 30 year SWR based on 120 years of stock market records.  There are only 3 completely independent 30 year periods in 119 years.  In that 120 years we had 2 world wars, the rise and fall of major industries,  the  rise of the USA from an emerging market into a world dominating economic super power and the destruction of global empires. What will the next 30 years bring?  Who knows but it could well be quite different to what has been seen before.

For example the recent large fall in bond prices is, I believe, unprecedented and only happened following an unprecedented rise over the previous 40 years.

So ISTM to make sense to derisk as far as you need to ensure your day to day income is protected up to the point where a projected scenario means the end of the world financial system as we know it.  Beyond that you just have to shrug your shoulders and accept you will have to join the queue for the soup kitchen along with everyone else.  This approach means that if your pension fund drops by 40% you know that your day to day living expenses  are safe for the next say 10 years and so you are under no pressure to act immediatly and end up doing something foolish.

More interestingly, what is your statistical grounds for saying that you cannot examine periods by month with overlapping scenarios?  Pretty much every model I've seen does that including some models that were developed by people who claim to have a degree in statistics - if you are saying their entire approach is invalid I'm surprised there isn't a bigger debate about it all the time?   I am not a statistician but I've seen some people write that the overlapping approach is valid as long as you use particular adjustments and methods in your statistics.

I am not a statistician either but I do use stats for long term risk assessments in engineering so have some understanding.

 Don’t think anyone said you can’t use comparatively short periods  taking month long or year long blocks and mixing them up differently to get a range of forecasts.  That’s what they do rather than “overlap”.  And in doing this mixing they may underestimate  “trends” and “momentum” scenarios which represent actual economic, fiscal and monetary conditions.  

But Linton’s point is valid.  They use this approach because they have no other option.  Its not great but you just don’t have a million years’ of data on stock markets.  This introduces all sorts of problems. For example if your “block size” is 1 year, the result would be very different to 10 year blocks (don’t think anyone uses 1 month because then you completely misrepresent “economic trends” and just get the average over 120 years).  In a perfect world we would have enough separate 40-50 year periods to do stat analysis but we dont.

Bottom line: there just isn’t enough available historical return data to run this type of experiment without getting creative.  If we had a million years of actual stock returns rather than just a century or so, it would be much easier and the estimate would have been more meaningful.  As it is, we have interesting simulations with questionable accuracy.  Certainly not the “95% confidence” some claim.  

Not sure what you mean by not measuring by month?  Several of the models I’ve seen do work by month.  What they do is take the current value of assets and convert it to a real terms starting point for each month in the last 100+ years.  

They then iterate through every month so you can see, for example, if you had retired in April 1973 with the same real terms situation you have now, would your pot have run out in x years.  You can then show how many of the months would have failed.  Not surprisingly it’s often particular months in the late 1960s or early 1970s or 1929 which are the last failing ones if you keep reducing your withdrawal rate.

Obviously all you can really conclude is whether your plan would have survived the worse market conditions on record in the last 100 years or so.  

Maybe I’m crazy but my feeling is that the guaranteed income types that you can invest in are in the end funded by the same economic capitalist system that you are relying on for your DC pot.  As such, any economic event that is much more catastrophic than anything seen since the 1890s is likely to cause solvency issues for the payment of even “guaranteed” assets and/or other major issues that will impact all of us anyway.

JohnWinder

‘Maybe I’m crazy but my feeling is that the guaranteed income types that you can invest in are in the end funded by the same economic capitalist system that you are relying on for your DC pot.  As such, any economic event that is much more catastrophic than anything seen since the 1890s is likely to cause solvency issues for the payment of even “guaranteed” assets and/or other major issues that will impact all of us anyway.’

Not crazy I’d say, but too charitable? We haven’t recently had an economic catastrophe, when some DB pensions got into trouble, that comes close to 1930 let alone ‘much more catastrophic’.

We well knew a century’s economic history when we offered people DB pensions, so they should have been realistic offerings. Or were they generous looking offerings to entice people into the job while accepting poorer wages in exchange for the deferred benefit of a compensating pension - never to be realised for some? The employer gets their cake and eats it by not funding the pension properly, or not be upfront about workable constraints on early withdrawal that can so damage long term investments, or not ending but grand-fathering some benefits which come to be seen as unsustainable . Sold a pup comes to mind with respect to a naive public. Is that overreach?

Linton

Pat38493 said:

Deleted_User said:

Pat38493 said:

Linton said:

Pat38493 said:

Mistermeaner said:

hi

question regards risk profiling

I understand that a number of investment companies / advisors will work with clients to understand their 'risk profile' and then structure their invetsments accordingly

i.e. someone 'risk adverse' would be put into generally lower volatility products such as bond weighted funds where as someone with a higher risk appetite would be higher in equities etc.

while i get the principle of this is it not somewhat self defeating as over a long term (esp with pensions) regardless of your 'risk tollerance' going bond heavy at age 20, 30 even 40 and 50 is simply lowering your returns and therefore exposing you to greater risks

Isn;t the concept just misplaced?

Reason for asking my company is moving to a new DCpension provider and the new company will be offering a 'risk review' to help people decide on their investments - i see the whole process as somewhat arbitrary and flawed

Appreciate thoughts and comments 


Thanks

As you can see from perusing these boards, a lot of people tend to get a bit of a panic on if they see their pension fund making precipitous drops in value, so if your pension fund dropped by 40% when you were one year into retirement, can you still stay rational enough to assume that this is still within the parameters of past history that you tested against, and therefore you don't need to take any action?  

Even if you know your strategy is fully rational, if you are the type of person who will lose a lot of sleep when your pension has dropped a lot early in retirement, I guess this is why people need profiling - you might be wealthier in the long run but you might not be happier if you are losing lot of sleep worrying about the recent market crash.

However fundamentally you are not far off - putting too much of your funds into low valatility and low risk investments is likely to end up with you working a lot longer.  

I know a couple of people who are fully aware of all this, but they still keep bringing up how much their fund dropped this year - go figure....

Rationally I suppose you should only start worrying if the markets do things that have never happened before and you were not prepared for in your retirement planning.

I think this is putting too much faith in the parameters of history.

In practice we have no really reliable parameters of history.  Take a 30 year SWR based on 120 years of stock market records.  There are only 3 completely independent 30 year periods in 119 years.  In that 120 years we had 2 world wars, the rise and fall of major industries,  the  rise of the USA from an emerging market into a world dominating economic super power and the destruction of global empires. What will the next 30 years bring?  Who knows but it could well be quite different to what has been seen before.

For example the recent large fall in bond prices is, I believe, unprecedented and only happened following an unprecedented rise over the previous 40 years.

So ISTM to make sense to derisk as far as you need to ensure your day to day income is protected up to the point where a projected scenario means the end of the world financial system as we know it.  Beyond that you just have to shrug your shoulders and accept you will have to join the queue for the soup kitchen along with everyone else.  This approach means that if your pension fund drops by 40% you know that your day to day living expenses  are safe for the next say 10 years and so you are under no pressure to act immediatly and end up doing something foolish.

More interestingly, what is your statistical grounds for saying that you cannot examine periods by month with overlapping scenarios?  Pretty much every model I've seen does that including some models that were developed by people who claim to have a degree in statistics - if you are saying their entire approach is invalid I'm surprised there isn't a bigger debate about it all the time?   I am not a statistician but I've seen some people write that the overlapping approach is valid as long as you use particular adjustments and methods in your statistics.

I am not a statistician either but I do use stats for long term risk assessments in engineering so have some understanding.

 Don’t think anyone said you can’t use comparatively short periods  taking month long or year long blocks and mixing them up differently to get a range of forecasts.  That’s what they do rather than “overlap”.  And in doing this mixing they may underestimate  “trends” and “momentum” scenarios which represent actual economic, fiscal and monetary conditions.  

But Linton’s point is valid.  They use this approach because they have no other option.  Its not great but you just don’t have a million years’ of data on stock markets.  This introduces all sorts of problems. For example if your “block size” is 1 year, the result would be very different to 10 year blocks (don’t think anyone uses 1 month because then you completely misrepresent “economic trends” and just get the average over 120 years).  In a perfect world we would have enough separate 40-50 year periods to do stat analysis but we dont.

Bottom line: there just isn’t enough available historical return data to run this type of experiment without getting creative.  If we had a million years of actual stock returns rather than just a century or so, it would be much easier and the estimate would have been more meaningful.  As it is, we have interesting simulations with questionable accuracy.  Certainly not the “95% confidence” some claim.  

Not sure what you mean by not measuring by month?  Several of the models I’ve seen do work by month.  What they do is take the current value of assets and convert it to a real terms starting point for each month in the last 100+ years.  

They then iterate through every month so you can see, for example, if you had retired in April 1973 with the same real terms situation you have now, would your pot have run out in x years.  You can then show how many of the months would have failed.  Not surprisingly it’s often particular months in the late 1960s or early 1970s or 1929 which are the last failing ones if you keep reducing your withdrawal rate.

Obviously all you can really conclude is whether your plan would have survived the worse market conditions on record in the last 100 years or so.  

Maybe I’m crazy but my feeling is that the guaranteed income types that you can invest in are in the end funded by the same economic capitalist system that you are relying on for your DC pot.  As such, any economic event that is much more catastrophic than anything seen since the 1890s is likely to cause solvency issues for the payment of even “guaranteed” assets and/or other major issues that will impact all of us anyway.

I think you nicely demonstrate my argument against taking SWR too seriously here.

All the failures in a high % success SWR  calculation occur from SOR in runs that happen to hit one of  2 specific major economic downturns. The critical simulation start times are relatively short. If either or both downturns had not happened or had been less severe the SWR could have been higher, if there had been 3 deeper falls the SWR would have been lower.  Is 2 separated by about 40 years normal in 130 years?  Were these downturns representative of what may happen in the future? What if 2 happened in a period of 10 years?  Who knows? We have no other data.  Having 130 years of data is irrelevent and does not add greatly to the accuracy of prediction. The only thing that may is a statistically significant number of 130 year periods.

In a previous post the impression was given that one should be blase about a 40% fall in a crash as the data shows that this is safe.  Sorry, the data does not prove this.

So I would suggest that SWRs are useful for sanity checking a retirement plan but of little relevence after that.

LHW99

Lies, d@mned lies and .....statistics........................................