Unobserved Components Model (Kalman Filter)

[Alexander]

Hi

#1 Im trying to forecast industrial production using Kalman filter. I’m following the paper by Antoniou et al. (footnote 1) http://web.bi.no:8000/MCFileManager/fil ... gp_jef.pdf where they use the unobserved components model:

Yt = Yt* + ut
Yt* = Yt-1* + et

I was wondering whether I am specifying the model correctly in EViews by using the following command:

@SIGNAL production = SV1 + [VAR=EXP(C(1))]
@STATE SV1 = SV1(-1) + [VAR=EXP(C(2))]

#2 Also, Priestley (one of the co-authors) in his previous paper (footnote 2) used the unobserved components model of the following form:

Yt = Yt* + ut
Yt* = Yt-1* + gt-1 + et

where gt = gt-1 + wt

Is the following command correct:

@SIGNAL production = SV1 + [VAR=EXP(C(1))]
@STATE SV1 = SV1(-1) + SV2(-1) + [VAR=EXP(C(2))]
@STATE SV2 = SV2(-1) + [VAR=EXP(C(3))]

#3 They also use the ‘autoregressive models with time varying parameters’ model of the form:

Yt = Xt.bt + ut
bt = bt-1 + et

Is the following input command correct:

@signal logcpi = sv1*logcpi(-1) + [var = exp(c(1))]
@state sv1 = sv1(-1) + [var = exp(c(2))]

Regards,

1 ’ Macroeconomic variables as common pervasive risk factors and the empirical content of the arbitrage pricing theory’ 1998 in Appendix A
2 The arbitrage pricing theory, macroeconomic and financial factors, and expectations generating processes, 1996, Richard Priestley pp.873
http://www.sciencedirect.com/science/ar ... 6695000356

[EViews Glenn]

From your brief description of the models, they appear to match the specifications.