Model combining SS and OLS equations

[dsinigaglia]

I am trying to use a forecasting model that combines several different types of equations. Essentially, the model has some OLS equations and identities (plain mechanic definitions). Now, I am trying to replace one of the OLS equations for an almost equivalent SS specification. The difference being the replacement of estimated coefs by random counterparts.

I have no problem with the specification of the SS equation or in estimating any of the equations. Actually, I can make a model adding up all equations, definitions (identities) and SS equation. The problem comes along when I try to make a forecast series based upon different scenarios for some exogenous variables. An error message appears stating eviews cannot recognize "SV_number" variable (the definition for the states in the SS equation).

Does anyone know if there is any restriction for combining OLS and SS equations + definitions within the same model?

An alternative would be to define everything under a singular SS model. The problem, in that case, is how to add up the definitions to the SS model since they establish contemporaneous relations of endogenous variables (which is not in line with SS representation).

Any hints? Thanks in advance.

[f.abbasi]

hi
Can I use lag length in OLS for this model?
Yt= c + yt-1 + c2x1t+ c3x2t+ c4x2t-1+ ma(1)