The seeds of the crisis were sown in the Great Moderation (the low inflation, relatively stable last 20 years or so). Everyone who patted themselves or others on the back during that time was really missing the point (p.3). The same interconnections that reduced the effects of small shocks created vulnerability to massive system-wide domino effects. No one saw this clearly.
This kind of model – in which greater resistance to small shocks can create vulnerability to large system-wide effects – has been employed to understand the relationship between reliability in electric power systems. It seems to be the case that at least many of the things that a local electric transmission system does to improve reliability work to push the larger system of interconnected local systems to a state in which it becomes more vulnerable to severe reliability failures – cascading blackouts. There seems to be a kind of frontier, given the current state of the transmission system, where we can choose to have more frequent small blackouts and a very low risk of a huge blackout, or we can choose to have infrequent small blackouts with a slightly higher risk of a huge blackout. (See, for example, work on cascading blackouts by Ian Dobson, Benjamin Carreras, David Newman and others, collected here, especially this paper.)
Of course, not every system shows this kind of interrelationship – making individual automobiles more reliable doesn’t increase the probability of a widespread automotive system failure, making telecom components more reliable doesn’t increase the probability of a cascading outage of phone services – and it is an open question whether this kind of model can be well employed to describe risks in the financial system.
To be fair, it is also an open question whether more conventional kinds of economic models can be well employed to describe the recent turns in the financial system.