Weakly exogenous variables - Cointegration

[MetricsLearner]

Hi there,

I have searched the forums and the manual and can't find an answer for this question, so any help would be much appreciated!

I have been using the Johansen procedure to test for multivariate cointegration between log of GDP (gdp), log of money supply (m) and the long-term interest rate (R) and have gone through the following steps:

- tested each variable for unit roots - all I(1)
- Found appropriate lag length for VAR including a structural break and seasonal dummies using information criteria - lag length 1
- Tested for reduced rank and used Pantula Principle to find that EViews model option 2 should be used for the cointegration tests (Intercept in CE not in VAR) and that the cointegration rank is 1 (and hence the cointegrating vector from the estimation is unique)
- Tested the alpha values for weak exogeneity and found that one of my three variables is weakly exogenous (R)

I am struggling to work out how to estimate the model where this variable is treated as weakly exogenous. That is:

Δy(t)= Γ(0)ΔR(t) + Γ(1)Δz(t-1)+ α(1)βz(t-1) +u(t)

where y(t) is the vector of the variables without R (including gdp and m) and z(t) is the vector including all three variables

I can't seem to find a way to include it solely on the right hand side as exogenous to the VAR/VECM and have lags/differences of the variable.

Any guidance would be really appreciated! I hope my question is clear enough. I have attached my workfile, I am using EViews 7.2