Page 1 of 1

Decision rules in deterministic models

PostPosted: Tue Jan 10, 2012 4:26 pm
by manof1kvoices
Hi everyone,

My question is very simple: why does not Dynare calculate the decision rules in deterministic models when the "simul" block is activated? it seems to me that it actually has to calculate them and store them somewhere in order to generate the time simulations for the different variables...am I wrong?

I read in the manual that decision rules are only stored for stochastic models using the "stoch_simul" block, but why in this case and not in the deterministic?

I would be interested in investigating the coefficients of the linearized decision rules for my variables.

Would greatly appreciate any help on this!

Thank you

Re: Decision rules in deterministic models

PostPosted: Mon Feb 13, 2012 4:09 pm
by SébastienVillemot
The reason is that there is no such thing as a decision rule in deterministic mode.

The concept of a decision rule is only relevant if the information set of the agent varies with time, and contains at every period a given set of past and present variables (the so-called state variables). The deterministic mode corresponds to a perfect foresight problem where there is only one information set (the one containing all the variables at all dates), so we cannot define a decision rule in this context.

What could be mathematically defined is the function which, given an initial and a terminal condition, and given a path for exogenous variables, returns the path of endogenous variables. This is precisely what "simul" does.