Emiel

As far as I know, you can run a regression with regressors that are originally of different integration orders, provided you first difference the I(1) variables, so that your regressions contains only I(0) variables. This is done and advocated for example (although slightly different method) in Juselius book on cointegrated VAR, where some variables are kept I(0).

Best

Matthieu

Le 10/08/2012 09:58, Emiel Denuwelaere a écrit :

Hello,

For my school these, i am doing research to investigate the effect of different variables on the reserve currency shares of different currencies. Before i am working with panel data, i liked to make an OLS estimation on the dollar. So i'm using time series data. I checked the variables for stationarity using the ADF-test, and it seems that the dependent variable is non-stationary along with two explanatory variables. The other 4 explanatory variables are stationary. Since you cannot run regressions with variables of a different integration, i am not sure how to proceed.. I suggest it is not appropriate to just take the first differences of the non-stationary variables?

Any help on this topic would be very much appreciated. Thanks in advance.
For additional info on the data/research, please do not hesitate to contact me.

--
Met vriendelijke groeten,

Emiel Denuwelaere

M: +32 (0)476 56 08 75
E: emiel.denuwelaere@gmail.com
_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users