I've been reading a lot of about the "Magnetar trade" (see pro-publica article here and the links therein, as well as this paper), and I'm slightly confused by the argument about how Magnetar (and similar firms) spurred the creation of additional CDOs, which might be related to a deeper misunderstanding of how the individual tranches related to the whole CDO.
So my understanding of the argument is that Magnetar bought the riskiest equity tranches of the CDO (thereby being known as the "sponsor"), which few investors wanted, which then opened them up to buying the higher rated senior or mezzanine tranches. This was great for Magnetar since they held short positions (CDSs) in those particular asset classes, and by buying the (undervalued) equity tranche they helped spur the creation of a larger CDO market in which to short.
What I'm unclear about is why these CDO's needed the equity tranche sponsored. Why was it not profitable (or seem profitable) for the banks that were creating these CDOs to make products with only mezzanine and senior tranches? Since it seems investors were buying individual tranches anyway, it isn't like the equity portion of the CDO was driving the return. I.e. those buying the higher tranches weren't necessarily being directly exposed to the equity tranche and their interest rate was based only on the assets in their particular tranche.