Is there an efficient frontier for including smaller companies alongside an index fund?

itwasntme001

And just to add, short term volatility and correlation (which impacts portfolio returns) parameters can change quite drastically.

Prism

itwasntme001 said:

How do we measure risk here?  Volatility over what period?

What about returns?  Over what period?

E.g. equities and bonds are negatively correlated over the short term, but long term they have been positively correlated for decades.

If you are truly long term investors, why do short term volatility and return metrics matter so much (which is what the efficient frontier is based on)?

Well as long we we compare like for like it shouldn't matter too much but both Morningstar and Trustnet produce 3 year standard deviation for their volatility risk score. I believe they both sample monthly.

Volatility doesn't matter day to day but everyone has their limit of what they are willing to accept along the way. Those in drawdown of their funds are more concerned as they are forced sellers. Reducing volatility is a researched benefit of a safer retirement where predictability of returns is something to be desired.

JohnWinder

Bonds cover a multitude of sins such that lumping them together can be tricky. And then there's 'long term'.

Portfolio visualiser shows the correlation of SP500 with intermediate Treasuries over 12 years is negative 0.32.

But to answer your question: why worry about volatility of you're in it for the long term? Because after a long term we hope we've turned bits of money each year into a lot of money, and so volatility causing a fall in value of 15% can cost you £150,000 if you had £M1. That matters. But perhaps we're no longer long term investors when we're well into retirement with £M1.

itwasntme001

Prism said:

itwasntme001 said:

How do we measure risk here?  Volatility over what period?

What about returns?  Over what period?

E.g. equities and bonds are negatively correlated over the short term, but long term they have been positively correlated for decades.

If you are truly long term investors, why do short term volatility and return metrics matter so much (which is what the efficient frontier is based on)?

Well as long we we compare like for like it shouldn't matter too much but both Morningstar and Trustnet produce 3 year standard deviation for their volatility risk score. I believe they both sample monthly.

Volatility doesn't matter day to day but everyone has their limit of what they are willing to accept along the way. Those in drawdown of their funds are more concerned as they are forced sellers. Reducing volatility is a researched benefit of a safer retirement where predictability of returns is something to be desired.

But volatilities can change, so not really sure of the benefits of using 3 year volatilities would be?  Especially when you are deciding between large cap stocks vs small cap stocks allocation.

Reducing volatility would probably be better done across asset classes (equities vs bonds) than within asset classes.

And I am not sure volatility captures all the risks.  It is a backward looking measure.  Just because "it is all we got" does not mean we should blindly follow it.

Prism

itwasntme001 said:

Prism said:

itwasntme001 said:

How do we measure risk here?  Volatility over what period?

What about returns?  Over what period?

E.g. equities and bonds are negatively correlated over the short term, but long term they have been positively correlated for decades.

If you are truly long term investors, why do short term volatility and return metrics matter so much (which is what the efficient frontier is based on)?

Well as long we we compare like for like it shouldn't matter too much but both Morningstar and Trustnet produce 3 year standard deviation for their volatility risk score. I believe they both sample monthly.

Volatility doesn't matter day to day but everyone has their limit of what they are willing to accept along the way. Those in drawdown of their funds are more concerned as they are forced sellers. Reducing volatility is a researched benefit of a safer retirement where predictability of returns is something to be desired.

But volatilities can change, so not really sure of the benefits of using 3 year volatilities would be?  Especially when you are deciding between large cap stocks vs small cap stocks allocation.

Reducing volatility would probably be better done across asset classes (equities vs bonds) than within asset classes.

And I am not sure volatility captures all the risks.  It is a backward looking measure.  Just because "it is all we got" does not mean we should blindly follow it.

They can change but don't seem to that much. When I first invested in Fundsmith one of the factors was lower volatility than many of its peers. It still has lower volatility. The same with most of my funds. Its just one of my criteria I look at and it can be measured. Most of the other measures of risk are abstract and much harder to pin down. What is the risk of manager failure? Or the risk of holding less stocks? How many stocks is required to mitigate the risk of single company failure? Its hard to put any kind of number on them or even find agreement.

Combining sectors is shown to reduce volatility historically. Combining large and small companies the same. It might not work every time and it might not even matter but its easy enough to do with little effort.

tebbins

maxsteam said:

rofl. It's not rocket science. If you want to know the risk of an investment, just look up the beta. Equities have a higher beta than bonds. Small companies have a higher beta than big companies. You take the risks that you are comfortable with. There's no "final frontier".

Beta and risk are not equivalent. There are many different types of risk.
What you're not getting is that, because equities and bonds are uncorrelated, holding both can be less volatile than a 100% bonds portfolio.

I think the mistake you're making is, for example:
Equity volatility 20%
Gilts volatility 10%
Therefore a 90% gilts 10% portfolio volatility is 90% x 10% = 9%, + 10% x 20% = 2%, total = 11%
However they are different assets, that behave differently and their correlation varies. Over most periods, introducing some equity into an all bonds (although it depends what kind) portfolio tends to reduce volatility up to a sweet spot around 20-30% equity.
There have also been decades like the 80s when UK equity was actually less volatile than gilts, and the 2000s when gilts outperformed equity.

aroominyork

jamesd said:

Prism said:

Here are Terry Smiths thoughts on the matter.

Fundsmith > Smithson > Financial Times - Busting the myths of investment

For convenience: "there is no doubt that adding a small/mid-cap element to a portfolio can achieve the seemingly impossible feat of generating additional return whilst reducing risk ... if you changed 35 per cent ... of your portfolio to MSCI World Small Cap you would get a higher return for the same risk. For any lower percentage, the higher return would be accompanied by lower risk".

And just for good measure, here is the efficient frontier graphic published in the Terry Smith article.

bostonerimus

aroominyork said:

jamesd said:

Prism said:

Here are Terry Smiths thoughts on the matter.

Fundsmith > Smithson > Financial Times - Busting the myths of investment

For convenience: "there is no doubt that adding a small/mid-cap element to a portfolio can achieve the seemingly impossible feat of generating additional return whilst reducing risk ... if you changed 35 per cent ... of your portfolio to MSCI World Small Cap you would get a higher return for the same risk. For any lower percentage, the higher return would be accompanied by lower risk".

And just for good measure, here is the efficient frontier graphic published in the Terry Smith article.

Interesting. This plot is to the upper right of the well known US based plot for various S&P500 and US bond allocations, just as you might expect. ie higher potential returns with bigger SDs. Many people will focus on the greater returns possible and forget about the increased variation about the mean of those returns.

aroominyork

bostonerimus said:

aroominyork said:

jamesd said:

Prism said:

Here are Terry Smiths thoughts on the matter.

Fundsmith > Smithson > Financial Times - Busting the myths of investment

For convenience: "there is no doubt that adding a small/mid-cap element to a portfolio can achieve the seemingly impossible feat of generating additional return whilst reducing risk ... if you changed 35 per cent ... of your portfolio to MSCI World Small Cap you would get a higher return for the same risk. For any lower percentage, the higher return would be accompanied by lower risk".

And just for good measure, here is the efficient frontier graphic published in the Terry Smith article.

Interesting. This plot is to the upper right of the well known US based plot for various S&P500 and US bond allocations, just as you might expect. ie higher potential returns with bigger SDs. Many people will focus on the greater returns possible and forget about the increased variation about the mean of those returns.

So by combining the two graphs - and using the historic average efficient frontier curves - we learn that the lowest risk combination of equities and bonds is 80% bonds, 16% large cap; 4% small cap. Hands up anyone who holds that?

bostonerimus

aroominyork said:

bostonerimus said:

aroominyork said:

jamesd said:

Prism said:

Here are Terry Smiths thoughts on the matter.

Fundsmith > Smithson > Financial Times - Busting the myths of investment

For convenience: "there is no doubt that adding a small/mid-cap element to a portfolio can achieve the seemingly impossible feat of generating additional return whilst reducing risk ... if you changed 35 per cent ... of your portfolio to MSCI World Small Cap you would get a higher return for the same risk. For any lower percentage, the higher return would be accompanied by lower risk".

And just for good measure, here is the efficient frontier graphic published in the Terry Smith article.

Interesting. This plot is to the upper right of the well known US based plot for various S&P500 and US bond allocations, just as you might expect. ie higher potential returns with bigger SDs. Many people will focus on the greater returns possible and forget about the increased variation about the mean of those returns.

So by combining the two graphs - and using the historic average efficient frontier curves - we learn that the lowest risk combination of equities and bonds is 80% bonds, 16% large cap; 4% small cap. Hands up anyone who holds that?

Maybe we need to update things for a low interest rate environment.