I was basically writing this exact same thing: economics as a science is a poor excuse for science (hat tip to Kedrosky... again.)
Economics is fascinating, because no one knows anything. Greider, in his excellent Secrets of the Temple said, in 1987:
The ultimate test for soundness for any science was the ability of its rules and theories to predict outcomes, and by that standard, economics was a crude and underdeveloped discipline.Twenty years later, still as true as it was then. Why are economics and medicine, where our ignorance daily cause huge hardship and lost lives, still relatively medieval in their development while physics and chemistry have been so dynamic? Why can we put people on the moon but we can't cure cancer? Why can we make fabrics that are breathable and waterproof but we can't mitigate the business cycle?
Economics has this virtue over medicine regarding the scientific method: there is a huge amount of experimental data being generated all the time. While it occasionally shows the theory to be correct (for example, Zimbabwe), it usually shows the theory to be completely wrong (for example, the, um, economy.)
The question is, why is this so? Inventing a better economic theory would provide much more in the nature of public or private rewards than, say, inventing a better explanation for why the universe is expanding at the rate it is. Discovering the Higgs boson may win you the Nobel prize, but being able to reliably predict the fluctuations of our GDP would give you some nifty trading profits.
In fact, you don't need to predict GDP very well, you just need to predict it better than the other guys. So, why doesn't somebody? "Cargo cults" aside, I can think of only one reason: we're just not smart enough. Or, same thing, it's just not possible. One thing's for sure: we need to admit how much we don't know. The essence of crisis management is minimizing the damage from low probability catastrophes. As we bounce from one of these to another, we should think about admitting that more than leaping without looking, we're actually blind. Building some slack into the system seems the least we can do.