You can probably assume people behave randomly

tags

Economics

In neoclassical economics agent rationality is central in the microfoundations of the supply & demande curve. Not only is the validty of this assumption is shattered by the Sonnenschein-Mantel-Debreu (SMD) theorem, it is probably not necessary. Unless we have a good reason to assume they are rational (it is a big assumption) then we'll have to do as if they were behaving randomly, hoping some emergent behavior appears and can be quantitied

What counts then, and what will shape the emergent behavior is the set of possibilities. If agents become correlated (bank run for instance) it is only the way the opportunity set is explored that changes, not the set itself. That would mean that At a macro level we should only be concerned with possibilities.