the problem is that commodities, as measured by the Reuters-CRB, for the period 1991-2004, returned less than T-Bills. I can't eat low correlation. Commodity prices are volatile, and if an investment in the asset classs is expected to achieve T-bill rates of return, wouldn't you be better off skipping the asset class (Commodities or whatever) and putting your money in something that rewards you for the risk?
The answer combines philosophy and faith. For me the philosophy is that by allocating to disparate asset classes I'll own the one that has a great year and there is less need to predict which one will be the one to have a great year. So if domestic stocks have a bad year it might be possible to bring up the overall return by getting a lot out of commodities or emerging market bonds or infrastructure or whatever. This has worked in the past and the faith part of the equation is that it will work in the future without having to rely too much on things "going right."
If you have eight different asset classes one of them will be the best performer, that best one may or may not be up a lot but often it is.
The example I used in my reply in the comments was market up 10% a diversified portfolio with 50 stocks being up 9-11% with one stock originally weighted at 2% being up 100% is very plausible. That one stock would account for a disproportionate amount of the total return.
It is is often the case that a disproportionate amount of a portfolio's return will come from just a few things. How is your portfolio YTD? Got any emerging market exposure? What is the weighting of that exposure and how much of your return came from that exposure? Do you have any energy stocks or an energy ETF? The energy sector SPDR is up 25% YTD. An equal weight to energy might be a big chunk of your return too. The Gold ETF (GLD), which is a client holding, is hardly an obscure holding is up 25% also. A 3% allocation from the beginning of the year would have added 75 basis points YTD which in an environment of SPX up 3.4% (plus dividend) seems like a lot.
The dilemma raised in Jimidean's question about commodities having trailed the return of t-bills during the period cited raises two points. One is that obviously that is in the past. The numbers might be a little different if he had picked a different 14 year period and more importantly looking forward are commodities likely to lag t-bills again?
The other point is that if he is worried that commodities will lag t-bills he should underweight the sector. If he is right, great but if he is wrong his portfolio can still get some benefit from an underweight exposure if commodities rocket higher. That is to say the consequence of being wrong could be lessened.
The view expressed in this post is just one of many valid ways to construct a portfolio and manage diversification. No one path can be the best path for all markets. My idea will work well some portion of the time as will more aggressive or more conservative approaches at other times.