However, I have read quite a few things about risk analysis in banks and there were lots of people that built models designed to calculate risk, probabilities of default, etc., and they used this to lower capital requirements and increase leverage.
One interesting thing I read (written by a quant) was that they built models that showed certain reasults. Then management came in and saw that the results didn't match their expectations, so they had the modelers tweak their models.
One of the biggest problems with the whole thing seems to be that the people making these huge financial decisions don't understand risk but thought they had the math down. See Fooled by Randomness, The Black Swan (both by Taleb), the Misbehavior of Markets (Mandelbrot), or one of the myriad of others talking about how the models just didn't match reality.