The problem with studying economics to predict human behavior: by the time you’ve learnt an “established theory”, the theory is obsolete. 
Yet today, unlike any other period in history, human behavior is actually quantifiable. The main flaw of economic theory – the asymmetry of information – has been mitigated by the internet (and creating as close to a perfect market as we can get).  Virtual economies provide measurable reliable data – something you don’t get in the real world. 
But then: A PROBLEM!
Measuring human behavior is done by taking large samples. Individual humans can vary tremendously, but take an average, and one “should” be able to account for “generic human behavior”, or at least, how any large group of humans should act.
This assumption goes down the drain when your large group stops acting like individuals, and behaves like a single entity.
Sadly, the availability of information does little to correct the asymmetry of information. By and large, like before the days of the internet, it seems that people are only privy to… what everyone around them is also privy to. Except where “everyone around them” used to be just limited to a small village or town, it can now encompass entire countries or regions.
Individuals don’t act rationally – that much is a given. But it used to be that a bunch of individuals, provided with the proper incentives, would respond to those incentives in some predictable, rational way. That doesn’t really work when the entire group behaves as irrationally as a single individual. 
Footnotes (pretty important points)
 – The paradox being that the same factor makes this both more and less true. The internet makes it easier for theories to become “established”, but it also rapidly changes human behavior.
 – Of course, some systems require such a depth of knowledge that they require specialists, which is slightly more difficult to model digitally. e.g. your robot lawyer is theoretically possible, but you’re not going to find 100 engineers who also have mastery of the law.
 – For an interesting follow-up, search for the time when Valve hired an Economics professor.
 – The point being: people seem to believe the first thing they read (and then re-affirm that believe). That’s great for survival in a world without mass media. It can be pretty horrible nowadays.