Robert Shiller, 18/3/2015
We must, nevertheless, acknowledge the irony of it all. We got into a mess because many people, many of whom were not economists, were playing with econometrics, making predictions outside of sample and assuming that the confidence that they had in the predictions was the same as if they were in-sample predictions. And because they were so damned wrong, the market crashed so hard that they fundamentally changed the behavior of the market. Now we are back to square one: we do not have enough data of this new behavior to learn how it works. We have computer power and tools, but not enough usable data.
The thing is, if you push the data in physics, even if you screw up, you're not going to fundamentally change the laws of the universe. If you push the data in economics, if you screw up, you can fundamentally change the behavior of the agents. In economics, you walk on egg shells, so it is best to slow down and be humble.