Permanant Portfolio With A Bias

If you look at any portfolio (or asset) compared to the correlation of the other ETFs (assets) in the portfolio, you can see that there will be an average of a slight positive or negative correlation to that aspect of the portfolio, even if overall the portfolio nets out to a relatively neutral bias (correlation average of zero). What this means is that you are going to be levered independently to a particular asset more than to your overall portfolio. Such "zero correlation" strategies aren't actually really "zero correlation". For starters the correlation is a look into the PAST, and you never know what they will be going forward. Secondly the independent variables may have a strong correlation while overall the portfolio zeros out.
For example.

	GLD	SPY	UUP
GLD	1.00
SPY	0.29	1.00
UUP	-0.30	-0.26	1.00

Although this has a slightly negative bias overall, it has a slighly positive one to the SPY. Thus if the SPY moves up or down, the rest of your portfolio will be slightly more likely than 50% to do the same (in aggregate). On the other hand, the dollar has a negative bias, meaning that if the dollar is up, the rest of your portfolio other than UUP is down.

	GLD	SPY	TLT
GLD	1.00
SPY	0.29	1.00
TLT	0.05	-0.41	1.00

A more extreme example. Overall this portfolio is very good but the TLT-SPY correlation creates a fairly moderate correlation against the SPY and much stronger against the TLT. GLD-SPY offsets some of the negative correlation but GLD-TLT hardly offsets anything. If TLT trends very strongly downward, the SPY is expected to trend upward strongly more often than not.

 It's not terrible to have this relationship. Afterall if ONE asset goes down strongly and the other up, you can sell the one that is up (reduce position) and BUY what goes down (add on). This is entirely how one can beat a "random walk" market.With a much flatter correlation, you reduce risks more because of the lack of exposure to any one area, or negative exposure.

25% of each 3, plus 25% cash beats a random walk or one in which prediction is impossible. However, if you don't believe in "efficient market" and think you have a significant edge, you can gain.

If it was just a choice between SPY and cash, if 50% represents a random walk (50% of increase, 50% of a decline), then 75% would be aggressively bullish (75% of an uptrend). You probably wouldn't be able to be too much more confident than 75% that you are near the bottom or in an uptrend over the next X years, nor will you be 75% confident that stocks will go down. As such you would position between 25% SPY and 75% SPY allocation. With more assets

However, If one was to observe that they believe they have an advantage and can predict a "bottom" in gold, rather than own say 25% of each (25% will remain in cash ready to rebalance when necessary), they may decide that they will start to accumulate a larger position as a percentage of the portfolio as gold goes more and more oversold. For example, even though gold is here around $1225 right now and anyone who previously held such a portfolio would be buying a bit more to rebalance since UUP and SPY are moderately up lately. If for example one held $500 in SPY and $500 in UUP and $500 in GLD, they might find their portfolio over the last 6 months, their portfolio has gone from

25% SPY, 25% GLD, 25% UUP, 25% cash to

about
29.30% SPY , 18.83% GLD and 26.38% UUP and 25.48% cash.

Rather than rebalance it, you may want to consider an alternative if you think gold is going to bottom soon, and perhaps SPY has been on a run and you are worried about it peaking soon.

So you shift from the 29.30% SPY , 18.83% GLD and 26.38% UUP and 25.48% cash.
TO
20% SPY 30% GLD 25%UUP 25% cash
Then say months later, GLD flushes out from $1225 to $950 and now the SPY and UUP goes significantly higher.
Now you add up even more because you are convinced this is the bottom in GLD and you really believe you have an edge. This strategy can be a bit dangerous, but if you really believe your odds and upside in the future have improved, you may rebalance to
20% SPY 35% GLD 25% UUP 20% Cash.

Now if from there, Gold goes up significantly, you are naturally going to have an even larger position and you will have bought fairly aggressively at $1225, and then aggressively again at $950. At this point you may want to rebalance back to 30% GLD or 25%. If it continues then you rebalance to 30%. And maybe TLT drops really low as UUP shoots higher and you want to sub out UUP for TLT.

The idea is you are positioning yourself to more aggressively "play the odds", and more aggressively "buy low" and "sell high" than simply believing in a "random walk". You have to really know what kind of market you are in and whether or not it is a correction, a bear market, or a long term decline like GOLD in 1980 or Nikkei in 1989. I don't think 2011 was a 1989 Nikkei or 1980 GLD type top. I Do think gold has been in a bear market. I am supremely confident that $950 in gold holds, but worried we could test that area, and even break below it temporarily (very short term). I could be wrong, but you have to be willing to take the risk in order to profit.

This has been my thoughts only, as always, not to be taken as investment advice. Full Disclosure: I bought SLV just prior to this post.