Skip to main content
Message Font: Serif | Sans-Serif
 
No. of Recommendations: 0
Let me start by saying that I don't believe that volatility is a proper measure of risk and this question is not about risk. However, this board seems the most appropriate place to ask this question.

I always believed that as a long term buyer and holder of a diversified portfolio of shares, the interim volatility doesn't matter. In other words, what matters is the value of my shares in 20 years' time when I will be cashing them for retirement. If there is a 50% stockmarket crash tomorrow, it's of no concern to me. I should invest in the asset class with the highest return and ignore the interim volatility.

Now let's say that I can invest in a choice of two portfolios for the next 3 years with the following returns:

Stocks: 50%, -41%, 50%
Property: 5%, 10% and 15%

An investment of $100 in either portfolio will rise to $ 133 after 3 years. Likewise an investment of $50 in both portfolios will also amount to £133 in total.

Now, consider what happens if I change my strategy so that at the start of each year, I rebalance my portfolio such that I have 50% of my portfolio in property and 50% in stocks. My portfolio will be much less volatile than the sole stocks portfolio as it will be evened out by property.

My annual returns are now 27.5%, -15.5% and 27.5%, so this mixed portfolio accumulates to $143 after 3 years.

I am leaving out the transaction costs in switching from property to equities and back again.

But my conclusion is that a more balanced portfolio which reduces the portfolio volatility will boost the overall returns.

Am I missing something ?

Brendan

Print the post Back To Top
No. of Recommendations: 0
you haven't missed anything. if two assets have the exact same long run total return and have negatively correlated short run returns, then the periodic rebalancing from one asset to the other will produce higher long run returns.

unfortunately, in real life there probably are not any two assets where both conditions are satisified over a long time period.

tr
Print the post Back To Top
No. of Recommendations: 0
Am I missing something ?

Transaction costs matter unless you have absurdly huge amounts of money to invest and don't care about taxes. [Or have a diverse array of NTF mutual funds in a retirement account.]

It's impossible to find negatively correlated assets in advance. If you knew the returns of each asset in advance, then year 1 you'd be 100% in stock, year 2 100% in property, year 3 100% in stock. If you don't know the returns of each asset for each year, you cannot determine their correlation.

You cannot "boost returns" unless you have a crystal ball or you increase your risk. Modern Portfolio Theory just makes it harder to see that you're making assumptions about the future.
Print the post Back To Top
No. of Recommendations: 0
I appreciate that I can't find two shares which are perfectly negatively correlated, but if I choose shares or asset classes which are not correlated at all, won't I achieve a partial effect ?

Brendan
Print the post Back To Top
No. of Recommendations: 0
You cannot know future correlation in advance, and it's future correlation that determines whether this strategy will succeed or just sock you with lots of transaction costs.

Most people who use this strategy assume that correlation doesn't change, i.e. the future will be the same as the past. However that's right about as often as attempting to predict future percentage returns based on what returns were in the past (not very).

For instance, stocks and bonds used to have a low correlation; now it is much much higher.
Print the post Back To Top
No. of Recommendations: 0
the problem you have with a strict program of periodic rebalancing is the potential risk of pulling your flowers to water your weeds.

what i'm saying is that aside from the tax and expense issues, such a program of strict capital rebalancing is more of a capital preservation strategy than a capital growth strategy.

tr
Print the post Back To Top
No. of Recommendations: 0
Can we make any reasonable assumptions about the future ?

Can we assume that a diversified portfolio of stocks will be correlated with a separate diversified portfolio of stocks ?

Can we assume that the return on two portfolios of stocks will be more closely correlated than the returns on a portfolio of stocks and say bonds ?

I have always just assumed that the best strategy is to be 100% in equities all the time and now I am not so sure.

My understanding of the central tenet of Modern Portfolio Theory that the volatility on a diverse portfolio of stocks is lower than the volatility of the individual component stocks ? Are you rejecting this ? Or are you rejecting CAPM, beta and subsequent developments ?

Brendan
Print the post Back To Top
No. of Recommendations: 0
Can we make any reasonable assumptions about the future ?
Those who have will act to preserve what they have, and the government will try to maximize revenues.

Can we assume that a diversified portfolio of stocks will be correlated with a separate diversified portfolio of stocks ?
How do you compose the two portfolios? How do you weight the stocks in each portfolio? I think there are too many variables to answer that question.

Can we assume that the return on two portfolios of stocks will be more closely correlated than the returns on a portfolio of stocks and say bonds ?
Here I can give a resounding NO. It might be the case some of the time, but it won't be the case all the time.

My understanding of the central tenet of Modern Portfolio Theory that the volatility on a diverse portfolio of stocks is lower than the volatility of the individual component stocks ?
Yes -- provided you compose the portfolio correctly. Also, without knowing all the correlations between the stocks, you won't know how much you lower the volatility.

A diversified portfolio of sufficient size will limit individual company risk.
Print the post Back To Top