In applying the cybernetic trading system to the futures portfolio, much of the system is the same as that used for the ETF portfolios (small and hedge fund), but there are some differences. Like the ETF portfolios, an asset allocation scheme is used. The asset classes are:
- US equities
- US treasury bonds
- International equities
Note that there is no real estate asset class. This is due to the lack of liquidity in the S&P/Case-Schiller Home Price Index futures contract.
Another difference is how trading capital is allocated among the asset classes. For the ETF portfolios, an equal $ amount was allocated to each asset class, with an added risk management rule. For the futures portfolio, $ portfolio risk is allocated equally to each asset class, thus risk management is explicitly built into the asset allocation scheme.
Due to the use of leverage and potential high volatility of futures (see cotton for an example), I have added a trailing stop with reentry. The idea is to lock in profits, while having enough “wiggle room” to stay in a trending market if the trend continues while volatility remains “reasonable”. A reentry rule allows the system to reenter a trending market after the trailing stop was triggered due to either a temporary pullback or high volatility.
Since futures contracts have finite lifetimes, there are two issues to consider. One is what type of continuation method to use to generate a long enough time series needed for the cybernetic system algorithm. Second is how to handle rollover.
For a large, meaning multimillion dollar, portfolio, I would closely replicate the hedge fund ETF portfolio. However, I wanted to use a small portfolio ($100,000) that would be accessible to a non-hedge fund investor. Thus, I have chosen to replicate the small ETF portfolio. Here are the futures used:
- US equities – ES – emini S&P500
- US treasury bonds – GE – eurodollars
- Currencies – YG – mini gold
- International equities – ER – emini MSCI EMI
The US equities and currencies asset classes are the same for the futures and small ETF portfolios.
For international equities, I have decided to use an emerging markets index for the same reason that I use VWO for the hedge fund portfolio. (Note: collective2 does not currently support trading for the emini MSCI EMI. I have asked them to add it. )
I decided to use Eurodollar futures for the US treasury bonds asset class for technical reasons. A quick look at long term bond charts shows that Eurodollars trend better and have less volatility than treasury bond and note futures. Thus Eurodollars are a better fit for the cybernetic system.
Commodities are problematic. There is no liquid futures contract for a broad commodities index. I decided to equally weight crude oil and soybeans as a proxy for a broader index. Crude oil and its derivatives usually comprise nearly half of most commodity indexes. As I have mentioned previously, I think that this over represents the importance of crude oil, but since I wanted to keep the number of holdings in this portfolio small, I was compelled to adopt this weighting. I chose soybeans as a proxy for all agricultural products due to their ubiquitous use.
Regarding the performance of the futures portfolio, I am mainly interested to see if the (annual rate of return)/(maximum drawdown) ratio is the same as for the ETF portfolios. This is of the utmost importance since it will allow me to select a maximum drawdown based on the amount of leverage used and I can use leverage without having to borrow from a broker.