Another 'feedback on pension fund allocation' please

Prism

Deleted_User said:

You need to compare like with like.  Once you properly  account for risk and compare over sufficiently long periods of time, the evidence that expensive funds do not provide better results becomes overwhelming. 

That seems to be the problem with these ongoing debates about fund choices and asset allocations. No one can agree on what risk is and if something is riskier than something else. For example is the World Low Vol index lower or higher risk than the World index? It has fewer stocks and higher concentration so higher risk? It has historically and likely going forwards had lower standard deviation and therefore has lower risk?

You also get risk scores that say that the US markets are lower risk than Europe and the UK. Is that really to be believed?

No one is going to agree on this stuff so I will leave the worrying about 1-10 risk scales, volatility, sharpe ratios etc to the financial lot who need to try and quantify it.

[Deleted User]

Prism said:

Deleted_User said:

You need to compare like with like.  Once you properly  account for risk and compare over sufficiently long periods of time, the evidence that expensive funds do not provide better results becomes overwhelming. 

That seems to be the problem with these ongoing debates about fund choices and asset allocations. No one can agree on what risk is and if something is riskier than something else. For example is the World Low Vol index lower or higher risk than the World index? It has fewer stocks and higher concentration so higher risk? It has historically and likely going forwards had lower standard deviation and therefore has lower risk?

You also get risk scores that say that the US markets are lower risk than Europe and the UK. Is that really to be believed?

No one is going to agree on this stuff so I will leave the worrying about 1-10 risk scales, volatility, sharpe ratios etc to the financial lot who need to try and quantify it.

True.  

This is where the whole “efficient frontier” theory fails.  To draw this line you need to assign expected returns, volatility and correlations.  It's bad enough to try to get an estimate of an individual standard deviation or of a correlation coefficient. Putting a number of these into a variance-covariance matrix and then inverting it is just asking for big trouble. The errors on each individual parameter compound and the result is garbage.

The easiest way to see this is to find the standard error associated with the "optimal" portfolio. It's difficult analytically, but a bootstrap should give you a pretty good estimate. It's very large.

Throw in the unfortunate fact that these parameters, and especially the correlation coefficients, change over time, and you get 2008 all over again.

For this reason, I would ignore the 1-10 scales, Sharpe ratios, etc.  What I wouldn’t ignore is the efficient market theory which tells us that in the developed markets the stocks are priced about right based on todays information. Given that, simple maths shows the power of diversification to achieve the same expected return at lower risk/volatility.  So, every time you put too much of your money into a single company, industry, currency, asset class or country, you are exposing yourself to additional risk.  Every time you are underweight in something - ditto.  Low vol funds concentrate on certain industries and thus reduce diversification and ADD risk. On top of that, once a factor becomes known and popular, money streams into it and the advantages disappear very fast. 

Own the world and you opimized your portfolio. Simple. Maximum diversification. Very cheap. No maths or MBA required.  

And if you fancy an alternative stratergy, read A Random Walk Down Wall Street before acting on it.