non-stationary data
Posted: Mon Dec 29, 2014 1:45 am
Hello,
I need help on non stationary data? Any help would be very much appreciated.
I'm using economic time series data to do regression model.
Y = c+b1X1+b2X2+b3B3+b4X4+b5X5+b6X6
1. I have done unit root test and found that some data series are stationary at first difference and some series are stationary at second difference.
2. Then I run regression on second difference but the P-value of all variables are higher than 0.05 - which I understand that all independent variables are not significant
3. However when I run regression at raw data (level) all variables are significant (but DW is very low and R2 is very high which I understand that it is spurious)
My questions are
Since it is found that all data are non-stationary. Although it is stationary at second difference , it is found that all independents variables are not significant.
1. How can I run regression model to forecast Y?
2. In case that second difference are stationary and when run regression, all independents variables are not significance. How can I get Y at level?
Thank you in advance for all helps.
I need help on non stationary data? Any help would be very much appreciated.
I'm using economic time series data to do regression model.
Y = c+b1X1+b2X2+b3B3+b4X4+b5X5+b6X6
1. I have done unit root test and found that some data series are stationary at first difference and some series are stationary at second difference.
2. Then I run regression on second difference but the P-value of all variables are higher than 0.05 - which I understand that all independent variables are not significant
3. However when I run regression at raw data (level) all variables are significant (but DW is very low and R2 is very high which I understand that it is spurious)
My questions are
Since it is found that all data are non-stationary. Although it is stationary at second difference , it is found that all independents variables are not significant.
1. How can I run regression model to forecast Y?
2. In case that second difference are stationary and when run regression, all independents variables are not significance. How can I get Y at level?
Thank you in advance for all helps.