I have the same two Matlab files as in the previous topic, my code works perfectly that way such that my agents are surprised by the evolution of exogenous variables at each date. But with a huge model, the Matlab files are not really user friendly. So I wondered if I could do the same just with one .mod file with the perfect foresight command and a CSV file for shocks values.
The new command perfect_foresight_with_expectation_errors_solver (and its associated shocks(learnt_in=…) and endval(learnt_in=…) blocks) can be used to simulate unexpected temporary or permanent shocks in a perfect foresight setup. So it implements a generalization of MIT shocks. You can do everything in one .mod file, without the need to use MATLAB code.