Open ShuyanHuang opened 5 years ago
Hello. In the formal model, agents are assumed to know the model parameters. If they do not, has to introduce a distinct description of how beliefs are formed. This will typically be formulated a la Bayes Rule, but one can take a behavioral approach. The 2015 JPE paper is precise about priors in these types of models. But the uncertainty agents face involves the choices of others, which are not observable, not uncertainty about social structure, etc. One can introduce uncertainty at that level as well. There is a literature on model uncertainty and choice. Hansen and Sargent is the richest treatment although in a completely different context.
Thank you for presenting! About your discrete choice models of social interactions, I am wondering how you incorporate the agent's uncertainty about the underlying model. Is there anything like a Bayes prior?