Main Page | Report this Page
 
   
Science Forum Index  »  Statistics - Math Forum  »  Please help me in eviews...
Page 1 of 1    
Author Message
...
Posted: Fri Jul 18, 2008 4:16 am
Guest
Hello. Does anyone can help me with this in eviews? I have the returns
and i want to make this model R(t)=c1+c2d2(t)+c3d3(t)+c4d4(t)+c5d5(t)
+$B-t(B Return(t-i)+e(t) where i=1...k are the lags. Can anyone tell me how
can I create $B-t(B Return(t-i) this in eviwes? Thank you
...
Posted: Sat Jul 19, 2008 12:38 am
Guest
On Jul 18, 3:16 pm, dimanagnostopou... at (no spam) gmail.com wrote:
Quote:
Hello. Does anyone can help me with this ineviews? I have the returns
and i want to make this model R(t)=c1+c2d2(t)+c3d3(t)+c4d4(t)+c5d5(t)
+$B-t(B Return(t-i)+e(t) where i=1...k are the lags. Can anyone tell me how
can I create $B-t(B Return(t-i) this in eviwes? Thank you

D
Are you sure this is the correct specification? This looks like an
unconstrained distributed lag model with some period dummies. In which
case there should be parameters on the lags e.g. $B-t(B a(i)Return(t-i). IN
practice you have to set the length of the lag a priori or undertake
some experimentation and relying on statistical criteria to determine
an 'optimal' lag length. Using the eviews command interface might be
as follows (assuming choice up to 8 lags) with fairly obvious
equivalent (the beauty of eviews) in tne interactive interface.:-
equation eq1.ls r c d2 d3 d4 d5 r(-1 to -Cool
For further details put 'lag' into the index search of the help
facility.
G
...
Posted: Sun Jul 20, 2008 2:50 pm
Guest
On Jul 19, 11:38 am, Gerald.Harb... at (no spam) googlemail.com wrote:
Quote:
On Jul 18, 3:16 pm, dimanagnostopou... at (no spam) gmail.com wrote:

Hello. Does anyone can help me with this ineviews? I have the returns
and i want to make this model R(t)=c1+c2d2(t)+c3d3(t)+c4d4(t)+c5d5(t)
+$B-t(B Return(t-i)+e(t) where i=1...k are the lags. Can anyone tell me how
can I create $B-t(B Return(t-i) this in eviwes? Thank you

D
Are you sure this is the correct specification? This looks like an
unconstrained distributed lag model with some period dummies. In which
case there should be parameters on the lags e.g. $B-t(B a(i)Return(t-i). IN
practice you have to set the length of the lag a priori or undertake
some experimentation and relying on statistical criteria to determine
an 'optimal' lag length. Using the eviews command interface might be
as follows (assuming choice up to 8 lags) with fairly obvious
equivalent (the beauty of eviews) in tne interactive interface.:-
equation eq1.ls r c d2 d3 d4 d5 r(-1 to -Cool
For further details put 'lag' into the index search of the help
facility.
G

Thanks you very much Gerald. You where very helpful.
 
Page 1 of 1       All times are GMT - 5 Hours
The time now is Sat Nov 22, 2008 6:17 pm