The changing conditions was the introduction of some rather clever mathematics by the backroom staff, they came up with fancy new ways of "assessing" risk and even better fancy ways of presenting those assessments, splitting up stuff that in the past no one thought about splitting up and so on.
This is one of several common assumptions, but I have seen no evidence of it, and am reasonably close to the issue, meaning I have access and long-standing interaction with these "backroom staff" you mention. I think "risk models" are pretty much the same ones that have always been used. These risk models break down when you need them most (such as in the presence of systemic externalities to the risk assumptions: evaporating liquidity, correlated mistakes). A lot of people
know that risk assessments are no good in these circumstances, but no "old" risk model in the past would have been any good either. Despite widespread knowledge that standard risk assessments do not protect you in a crisis, they still get used because crises are not normally predicted (or predictable) by definition. In short, there really isn't anything that was done differently on the risk assessment/presentation side.
This all worked very well on paper and I bet the backroom staff understood what was really being traded, unfortunately these weren't really understood by the front men, and I say this was endemic from the CEO down to the interns, in other words they hadn't a clue what underlay these new schemes. Given this they simple failed to introduce new due diligence ideas alongside the new trading schemes.
This statement seems to be the other one I have heard repeated most often which is "there was not enough transparency in this stuff; only the boffins could understand it", and again I have not seen the evidence of it. The mechanics of the possible financial outcomes of leveraged securitised debt positions are
not, actually rocket science at all, and I think the front people know/knew what they are (I have not heard Dick Fuld or anyone claim: "I did not know what the risks were", which of course would be unforgivable). But in particular I think it is
not the case that the pricing of many such securities melted down
because "nobody understands them". I think it is more accurate to say that this happened because everyone
did (more or less) understand but
they all bet wrong in a correlated fashion, which in itself rendered "risk models" invalid. In this respect the primary thing that was not apprehended was
the extent to which everyone had a leveraged bet the same way round.
This is a common feature of many boom-crisis episodes, and I tend to believe that it is also its own
reason why one sees more boom-crisis episodes than
should be the case on the assessment of "risk models"
This is why I agree with those who suggest that by far the most important policy fix is one that regulates leverage, above anything that calls for still-fancier risk assessment and/or greater transparency. I think that both of the latter largely bark up the wrong tree. (Of course right now nobody needs a regulation against "excess" leverage--but that will not be true for ever)
