State-space modelling in Eviews 9.0

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

bgauchan
Posts: 13
Joined: Tue Aug 15, 2017 7:56 am

State-space modelling in Eviews 9.0

Hi,

Appreciate if you can help in defining state-space model (with kalman filter) for determining common component of structure shocks u1 and u2 corresponding to gdp_growth and gdp_defaltor respectively for few countries under consideration for comparison purpose. The structural shocks were arrived at after running svar. The idea is that if the proportion of common component is higher across the countries by running state space model we can conclude that they face symmetrical shocks and are suitable for common currency. The file is attached for your perusal.

1. I have tried for Bangladesh (ss01, ss02) as well as all the countries together using supply shocks sheet (ss01, ss02) but it doesn't seem to work. I am also not sure if it has to be done individually for each country or jointly for all the countries together (in which case there is separate sheet corresponding to each of supply shocks and demand shocks). KIndly advise the right process;

2. Appreciate if you can assist me by creating state space model as appropriate so that I can replicate it for the rest.

3. Which is common component and country-specific component and how to measure them and compare across countries?

Attachments
oca_bivariate_sspace - copy.wf1

bgauchan
Posts: 13
Joined: Tue Aug 15, 2017 7:56 am

Re: State-space modelling in Eviews 9.0

Following Chemie, DeSerres and Lalonde (1994), state-space model is used to decompose structural shocks into unobserved common and country specific components to estimate symmetries of shocks across the countries in the region. The screen shot attached herewith shows the results of the following state space model for supply shocks:

bgd = c(1)*s+[var=c(2)^2]
ind = c(3)*s+[var=c(4)^2]
npl = c(5)*s+[var=c(6)^2]
lka = c(7)*s+[var=c(8)^2]
@state s=c(9)*s(-1)+[var=1]
@param c(9) 0

As you can see, the results do not show standard errors, z-statistics and probability corresponding to each coefficient. It seems that the above state space model is incorrectly defined. Please advise.
Attachments
Chamie et al(1994) OCA Europe&US.pdf
oca_bivariate_sspace - copy.wf1
Screenshot (29).png (157.89 KiB) Viewed 449 times

bgauchan
Posts: 13
Joined: Tue Aug 15, 2017 7:56 am

Re: State-space modelling in Eviews 9.0

I AM STILL WAITING FOR THE GUIDANCE....