Before the crisis, securitization of debt was quite uniformly seen as a very good idea, after all it made house ownership available to many families. After the crisis the assessment is much more negative, in fact there is a large backlash against the idea, seeing at the root of all evil. Of course, the truth is somewhere in the middle. Securitization provides a powerful way to diversify risk, but as the crisis showed, it can make fraud easier.
Guillaume Plantin tries sort this out by looking at how banks react to the availability of securitization. He points out that the risk diversification gives less incentives to banks to screen well loans. This is not necessarily a bad thing, as the optimal contract literature would tell you that the bank would be required to bear some more risk. The fact that their would be more borrower failures would have to be weighted against the increased availability of credit, and society would overall likely be a winner from securitization. The problem is that these contracts were not optimal. There is evidence that banks have been negligent if not misleading with information about the underlying loans. The issue thus goes beyond the (fixable) moral hazard with selecting loans. Indeed, when banks have private information about the loans, we have a lemons problem wherein they push the worst loans to the securitization market (and when the US Treasury buys up those loans from the banks, of course it inherits the lemons as well). The policy prescription is clear: either restrict to some degree securitization or, better, alleviate the opaqueness of the securitization market.