Tuesday, May 30, 2017

Blending factors in managed futures

Investors are often blending managers to find the right return to risk trade-off. The portfolio structuring is often simple. The correlation across managers is analyzed on a quantitative basis and descriptions of styles are used for qualitative analysis. This may work but often leaves a lot on the table for finding a mix that will be effective in the future.

Another approach which moves to a deeper level is to mix risk factors. Using a simple approach of looking at blending asset class factors, an investors can breakdown the performance of a CTA into asset class betas; an equity, fixed income, commodity and currency market risk. The blending managers can be run through a simple beta analysis to find managers that either enhance of neutralize the beta exposures of any individual managers. Of course, this approach does not count for the dynamic nature of market betas for a manager across time, but it can serve as a simple check on what risks are being taken with a portfolio of managers. 

In the above case, we have broken down the performance of two managers using monthly data over the last three years. We are simplifying the discussion and not including the standard errors on the betas. In the case of CTA #1, we are looking at a longer-term multi-model systematic trend-following manager. Notice that the manager has a low beta to the stocks market but a song beta to long-bond. There is no sensitivity to commodities and a slightly positive beta to the dollar. This is a good diversifier, but an investor may want to further decrease the sensitivity to traditional assets. 

In the case of further diversification, we can add CTA #2 which has a short-term focus. Here there is no sensitivity to the equity market portfolio. The only positive relationship is with bonds but it is lower than the exposure seen with CTA #1. CTA #2 also has a positive alpha component.  

Adding CTA #2 to the first manager will be able to neutralize some of the beta exposures to traditional asset classes. This work could be seen through simple correlation but the added step of looking at betas will further refine what is happening to a portfolio mix. This added step is very easy to add and allows for a deeper understanding of potential risks.

No comments: