Game theory is increasingly getting better*. For example, they model learning effects (rather than assume equilibrium). They still assume neoclassical “rationality” though.
I think research on “market” outcomes with regards to voting are a good example of where game theory is at. The boundaries are still set (laws), but within that framework the models allow for a lot of decentralization/chaos.
- I am basing this post on the game theory I had in college and a friend of mine doing a phd on rational voting (ahum). Maybe there are areas I overlook though…