Help with Dynamic Factor index

[mkgunasinghe]

Hi all,

I am hoping to create a common factor applying a Kalman filter on four monthly indicators using EViews. My state space form involves two steps. Any help/comments are much appreciated.

First, my measurement equation links the observable four indicators, y(i=1...4,t), to the common factor, c(t), as follows:
y(i,t) = A(i) * c(t) + e(i,t)

A(i) is the indicator specific factor loading that relates the common factor to each observable indicator, and error term reflect idiosyncratic movements.

The second step describe the dynamics between the common factor and residuals of the above measurement equation as follows:
c(t) = B * c(t-1) + v(t) where v(t) is iid N(0,1)
e(i,t) = H(i) * e(i, t-1) + n(i,t) where n(i,t,) is iid N(0, s(i))

This means that both the common factor and residuals are AR processes. So far I got this on eviews:

@signal cz_bci = c(1)*sv1 + e1
@signal cz_cci = c(2)*sv1 + e2
@signal cz_ip = c(3)*sv1+ e3
@signal ge_ifobus = c(4)*sv1 + e4
@state sv1 = c(13) * sv1(-1) + [var=exp(c(14))]
@ename e1
@ename e2
@ename e3
@ename e4
@evar var(e1) = c(5)*e1(-1) + var=exp(c(9))
@evar var(e2) = c(6)*e2(-1) + var=exp(c(10))
@evar var(e3 )= c(7)*e3(-1) + var=exp(c(11))
@evar var(e4) = c(8)*e4(-1) + var=exp(c(12))

My four signals are the four observable variables which consist of a indicator specific constant, the factor and an error term. The state sv1 is the factor with its autoregressive component and error term that is iid N(0,1). I also define the observable variable specific error terms using the evar var(e1..e4) equations.

When I do Proc -> Estimate I get this error: Syntax error in specification for equation "@EVAR VAR(E1) =
C(5)*E1(-1) + VAR=EXP(C(9))".

I would much appreciate the help of someone much smarter than me. Thank you very much, Milo

[mkgunasinghe]

I am using EViews 9SV. Apologies for not mentioning this before!