I found this from Alphaville to be very interesting. It goes on to explain how commodity etf growth is a function of the need to offload commodity exposure by the large financials. Sort of like creation of CDOs was driven by the need to offload housing exposure and desire to short the whole thing by opportunistic hf managers. Of course the retail public ends up holding the bag yet again...
ETFs have been a useful tool to allow banks to move risk off balance sheet. When a bank takes on risk through lending to or financing a big commodity player, say for an acquisition, there is a need to hedge potentially huge commodity exposure — so as sell to lock in the commodity price, and you couldn’t sell that volume easily into the terminal market; although you could transfer a large amount of exposure to investors through an ETF more easily. The only way this used to be done is by the bank taking proprietary risk, but they now have other risk issues and aren’t prepared to carry that sort of exposure.
Using ETFs becomes a mutuality of interest, with everyone moving to launch products to investors – retail and institutional – so that they can carry the risk instead. It’s all about de-risking your book. And you saw it in the dot com bust when investment banks pushed dotcoms, but offloaded the risk to investors. When the NASDAQ crashed the investors carried the bulk of the exposure.It all has similarities to the Abacus CDO. If you want to short the market you have to create the demand, like Paulson did. If you are an institutional advisor you can do that by hyping the commodity.
The rest can be found here http://ftalphaville.ft.com/blog/2011/05/11/565941/if-we-build-it-they-will-come/