Here's a point of view that's illustrates nicely, I think, several of the points that I've been trying to make over the past few weeks - nicely applied to the real world. We have Professor of Marketing and Statistics at Penn State that is advocating the prying open of large businesses to get accurate information and the use that information to model some economics that actually work. He rightly points out that where we've had mathematicians and scientists making models before, but with companies sitting on their data as tightly as possible, these models have been flawed from the beginning because they haven't had accurate real world data on which to build. Economics is a horribly complex system (not as bad as the weather or biology, but still pretty bad) and we're a long way from being able to model it perfectly but with the right information used and shared properly we can get a pretty good idea about how to handle certain types of situations and avoid some of the bigger risks.
This is the thing about models. They're not perfect. If they were it wouldn't be a model, it would be a replica. they help us understand what is actually going on rather than what we want to think is happening though. And if they're not based on realistic assumptions, they will fail.