Re: cointegration

On Mar 28, 12:08 pm, joseph Frank <josephFrank1...@xxxxxxxxxxx> wrote:
Hi,

i have a linear regression model and i have received a remark to check the distribution of the dependent variable to see whether there is a cointegration problem. Can anyone tell me what does this mean?
thx

If x and y have "unit roots" but the residuals e(t) of the regression

y(t) = a + b*x(t) + e(t)

are stationary, then x and y may be cointegrated. So you need software
that tests for a unit root vs. stationarity. R (including its
packages) and other software has this, as well as the ability to fit
cointegration models.

.