[Gretl-users] Time varying "latent" variable estimation via the Kalman filter

[Luke Hartigan]

Hi all,

I am interested in estimating an unobservable parameter via the Kalman
Filter using Gretl.

Basically, I need some help in how to formulate and estimate a time varying
NAIRU model:

Measurement equation:
Inflation_t = B_1*Inflation_t-1 + B_2 * (Unemployment_rate_t - NAIRU_t) +
e_t

Transition equation:
NAIRU_t = NAIRU_t-1 + n_t

The part I am unsure of involves specifying the model in state-space form.

I have consulted the manual and while it is very useful, I just need a
little more guidance if possible.

Thank you

Luke

Title: Time varying "latent" variable estimation via the Kalman filter

Hi all,

I am interested in estimating an unobservable parameter via the Kalman Filter using Gretl.

Basically, I need some help in how to formulate and estimate a time varying NAIRU model:

Measurement equation:

Inflation_t = B_1*Inflation_t-1 + B_2 * (Unemployment_rate_t – NAIRU_t) + e_t

Transition equation:

NAIRU_t = NAIRU_t-1 + n_t

The part I am unsure of involves specifying the model in state-space form.

I have consulted the manual and while it is very useful, I just need a little more guidance if possible.

Thank you

Luke