Page 1 of 1

Dummy variable problem in my model in differences

Posted: Wed Jul 07, 2010 1:25 am
by lugoga
Hello,

I’m estimating a vector error correction model with a dummy variable with the following specification.

year value
1970 1
1971 -1
1994 1
1995 -1
1996 1
2001 -1
2007 1
other 0

The model has pretty god result but I’d like to test it with 7 dummies, one for each of the years mentioned. My have data from 1968 to 2008 so I am not including one dummy for each year of the sample. The dummies take values different from cero in these years.

Dummy Year
D1 1970=1
D2 1971=-1
D3 1994=1
D4 1995=-1
D5 1996=1
D6 2001=-1
D7 2007=1


I get a near singular matrix when estimating the model in differences. I solve it dropping the first dummy but that is not what I want becasu I want to know the value of that coefficient. I can not solve it bt dropping another variable, just the first one.

Can someone please give me some help.

Thanks in advance,