The Seductive Nature of Return Distributions!

Our guest in this week’s episode of the Systematic Investor podcast series was Jerry Parker – our resident legendary Turtle – who, in his typical style, voiced his strong opinion on multiple topics we discussed.

Now, we are not going to provide you with a laundry list of these topics as we hope you would tune in and find out yourself. Instead, we would like to spend a couple of minutes discussing return distributions.

Don’t run away please – this is not going to be a boring statistics lecture. It will be highly relevant, especially for those of you who are asset allocators of large pools of money.

A return distribution is nothing more than the stream of returns that an asset provides to its owner. Single securities (like individual stocks and bonds) provide return streams but so do whole asset classes (like equity, fixed income, commodities) and investment strategies (Buy & Hold, Long-only benchmarked, Trend Following, Long-Short volatility, etc).

A return distribution is characterized by its mean, standard deviation, skew, and kurtosis – the four most important but not exclusive distribution moments. In previous blog posts, we already discussed the skew properties of Trend Following (positive skew) vs Long-Only stock and bond portfolios (negative skew), and the portfolio-level benefits associated with this.

Therefore, today we will address the first moment – the mean – in more detail.

“Just because the S&P500 has delivered 7-8% annual return over the last 100 years, doesn’t mean that it will continue to do so over the next 100 years” – this is a quote from this week’s episode and it zeroes into the problem associated with the mean of a return distribution.

You see, people assume that with a large enough sample, they can reliably estimate and predict the future expected returns of – in this case – S&P500. This statement is – at best – inaccurate because the mean of a return distribution can exhibit significant time variation due to a myriad of factors such as macroeconomic conditions.

Put more simply, there may be periods when the expected return of a strategy can shift to a significantly lower level than what has been observed historically as Ilmanen et. al. prove in a recent award-winning study.

You may believe you have superior forecasting abilities and you will be able to predict the future economic environment. You may be thinking that you’ll be able to forecast how the different strategies and asset classes in your portfolio will perform and you will be able to successfully rotate away from the losers? Well, time and time again empirical studies provide evidence that market timing is a very low Sharpe Ratio trade, and the reason for this is that forecasting expected returns is very, very difficult as another great friend of TTU – Rob Carver – writes in his book Smart Portfolios.

So, where does that leave us? What should we do if we know that the mean of the return distribution can shift and we know we can’t predict that?

Well, as usual, the best approach is to seek more diversification. In a world where the expected return of U.S. and global equities (which dominate the risk profile of most investor portfolios) can be vastly lower than what we have seen in the past, more than ever what we need is a time-tested uncorrelated strategy with a positive skew that is liquid and scalable.

“Does such strategy exist?”– you may be thinking. Let us give you a hint – it does - its name starts with “Trend” and ends with “Following”.

We hope you enjoy this week’s episode and as usual, we promise your spend your time well listening to it!