Thursday, November 28, 2013

Money Honey Portfolio Proposed Future Improvements

As the sole hedge fund manager of my own private equity (hur hur), I've been floating a lot of ideas in my head recently, only how I can improve my asset allocation model with the aim of adding alpha, reducing volatility and also simplifying the process for myself.

My first idea is implementing a market timing strategy to my portfolio. Mebane Faber has written an excellent paper here which highlights the difference in including a simple form of trend following market timing. He uses the simple 10 month moving average. Page 40 is the really interesting page.

The second idea that has come to mind is to allocate a scoring system based on 2 moving averages and price. By scoring all the current asset classes based on this system, the portfolio can be mechanically defined to OW or UW an asset classes. This idea came to me from 2 sources. The first source was an article that I just can't seem to find or remember. I think it was from Seeking Alpha, and the author wrote about how you can use moving averages based on their crossover and slope to decide when and how much allocation you would give. I have found a similar article by Chris Ciovacco here and here, who incidentally is the influence of my second source, which is introduces the idea of scoring. I think he does this in a very big and comprehensive way with many indicators with the model his company uses. I have a nagging feeling that he actually wrote the article that I was talking about!

Anyway, reading the first paper has given me quite a bit of insight towards using this asset allocation model. Perhaps other than junk bonds and EM bonds, which are the most volatile, the other bond classes do not seem to be largely affected in terms of both returns and volatility from the market timing model. I would imagine it is because they are both greatly muted in bonds comparatively anyway. Therefore, it might be prudent to instead tweak holdings in a minor fashion instead of eliminating or greatly reducing them. This will help anchor the portfolio by ensuring that a large portion of allocation is not suddenly shifted to equities, so as to dampen any shocks in the system.

I have to slowly come up with a system and review moving averages in relation to price before I can come up with anything more concrete. Imagine if this because a fantastic, simple DIY system that people can adopt, add/remove asset classes, monitor themselves and produce good risk-adjusted returns. I would be quite happy if that was the case!

I think the first step of the day however is to start recording down daily closing prices of all the funds that I track. Even though my data will be calculated from daily data (the most sensitive and accurate), I will most likely be looking towards weekly updates of my indicators, their interpretation and implications towards my monthly overview. My monthly overview will take into account the changes in weekly data, though I would like to keep portfolio rebalancing to a minimum, therefore the monthly implication must cross a certain threshold in value, or technical significance to warrant me to rebalance earlier than planned.

No comments:

Post a Comment

Observe the house rules.