inflation expectations as observable

[lizzie3108]

Hi everyone!

I have been reading the paper of Del Negro and Eusepi (Fitting Observed inflation expectations) and I came out with a question maybe a silly one but I prefer that someone who knows Dynare better can help me. In this paper they say thet they make inflation expectations observables when they make the DSGE estimation (a bayesian one), how is it possible? Is it possible to make pi(+1) observable? or there are other ways to do it? Please if someone could answer this it would be really great.

Thank you so much

[glucke]

It's really easy. You just declar additional variable e_pi and write additional equation: e_pi=pi(+1);. After that you have usual observed variable e_pi (I think that it is better to add measurement errors for this variable (I don't check was it in the model or not)).

[lizzie3108]

ok thanks, but how do i get measurement errors? and why should i include? sorry for these equations but I am not so good at dynare.

[jpfeifer]

by defining an additional shock, e.g.

Code: Select all

varexo eps_pi_exp;


and then using

Code: Select all

e_pi=pi(+1) +  eps_pi_exp;

and later on

Code: Select all

varobs e_pi;

He suggests this because inflation expectations are usually very poorly measured and adding measurement error thus appears sensible to improve the fit and reduce estimation problems.