Re: [Gretl-users] No lags possible
On Mon, 13 Aug 2012, Emiel Denuwelaere wrote:
I am trying to regress some time series variables in Gretl. The data
i'm
using covers 16 time intervals on 6 explanatory variables for one country.
When checking for autocorrelation i found the following critical values:
dL=0,615 and dU=2,157 (http://www.stanford.edu/~clint/bench/dw05a.htm). The
DW statistic from the OLS estimation was 2,477 so i found the Durbin-Watson
test to be inconclusive.
The test is inconclusive if the statistic falls between dL and
dU; here it doesn't.
I would be very thankful if anyone could help me
with the following two questions:
1) Do i need to use normal OLS or build a dynamic model?
2) I tried to build a dynamic model with lags, but i could only add 1 lag
from each variable. Gretl doesn't allow me to add more lags.. Is this due
to the short data sample?
6 regressors with two lags + current values = 18. Observations
= 16. You can figure it out.
Allin Cottrell