Hi Bob
On the basis that stocks generally provide the better longer term investment reward then the ideal combination of total holdings is where the individual classes each achieve around the same investment reward as stocks over the longer term - but do so via different price path motions (non correlated).
When over the longer term any single class under-performs the longer term stock average then that individual class drags down the whole.
Cash generally acts as a drag due to its longer term relative under-performance of stocks. Equally inverse ETF's generally also fail the longer term investment return objective.
Over shorter periods however cash can/does relatively outperform stocks, equally so can/do inverse ETF's. In many respects switching some funds into a inverse it little different to that of selling more stock and moving into cash. If you're long stocks 60%, 10% in cash and 30% in inverse ETF then that's comparable to being 30% in stocks, 70% in cash.
I have no idea as to whether when the boundaries were reached, such as reaching perhaps 10% long stocks, 10% cash and 80% inverse ETF (equivalent to 70% inverse ETF and 30% cash) would be rewarding based upon AIM indicated trading action as I've never tested such.
If you're seeking B&H beating returns over the longer term then I believe a better approach is to use 100% virtual cash across all AIM accounts and diversify across a range of asset classes that exhibit equity like returns over the longer term. An example might be to use four AIM's, perhaps one each for Stocks, Commodities, REIT's and Managed Futures, with each AIM assuming 100% of the total fund was allocated to it and with 25% stock, 75% (virtual) cash initial settings (or settings more appropriately selected according to how each asset class had recently performed).
A complication here is that you ideally need one class to be selling as another is buying, and in approximately equal capital amounts. This problem might be reduced by when there were more sells than buys you created some actual cash holdings, and when there were more buys than sells then any cash reserves previously built up could be deployed.
A risk is that any one class that was slowly failing could drag down the others. Such risk could be reduced by limiting the amount of stock value any one AIM held to a maximum proportion of the total fund value (perhaps 33%).
JMO
Regards. Clive
News 
Market Data 
Discover 
AI
