# st: problem with VEC (vector error correction models) and exogenous variables

 From "Vasilis Sarafidis" <[email protected]> To <[email protected]> Subject st: problem with VEC (vector error correction models) and exogenous variables Date Tue, 23 Jan 2007 15:01:58 +1100

```Dear Stata list,

I am trying to estimate a vector error correction model but I face the
following problem:

Let y(t) be a column vector of 2 endogenous variables, [y1(t) y2(t)]',
and let x(t) be an exogenous covariate.  Also, let z(t) be a column
vector of all the variables, i.e. z(t) = [y1(t) y2(t) x(t)]'

I want to estimate the following VECM (D is the first-difference
operator):

Dy(t) = B0 * Dx(t) + B1 *  Dz(t-1) + (a*b')*z(t-1)

so effectively one exogenous variable appears in the cointegrating
vector.

I checked the manual for the command VEC but I couldn't figure out how
to estimate this model.  One solution could be to treat x(t) as
endogenous, restrict the speed of adjustment parameter of x(t) to 0 and
impose the restrictions that the short-run parameters in the x(t)
equation are all 0.  But again, I couldn't see how to impose constraints
with regards to the short-run parameters.

Any help is highly appreciated.

Vasilis

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```