VAR problems!

[Jeanvion]

Hi guys,

I despirately need your help! My paper needs to be turned in next monday and i still need to analyse a relationship. If i can't finish this off tomorrow i won't be able to graduate this year... :? So to everybody who is prepared to help me out, thanks a lot! I am a newbie to eviews and know only the absolute basics. I use the fifth version of Eviews.

One of the topics of my paper researches the relationship between soccer attendencies (FANS) and competitive balance. My research is based on another paper i found named "An empirical analysis of competitive balance (CB) in European soccer leagues" from Leif Brandes. He used the VAR model to research the relationship. I already have all of my data and made it stationairy with 1ste differences, so i dont have to worry about that any more. So i want to know the relationship between Fans and each CB seperatly. For example the relationship between Fans and STD, Fans and CR, Fans and H-index.

In his paper he builds the estimation results in equations. In the paper the corresponding VAR specifications can be expressed as:

〖CB〗_t= α_10 + ∑_(i=1)^j〖β_1i 〖CB〗_(t-1) + ∑_(i=1)^j〖γ_1i 〖Fans〗_(t-i) 〗〗

and

〖Fans〗_t= α_20+ ∑_(i=1)^j〖β_2i 〖CB〗_(t-1)+ ∑_(i=1)^j〖γ_2i 〖Fans〗_(t-i) 〗〗

" Where CB_t-i i=1, ..., j denotes the value of the competitive balance measures for the season [t-1, t-(i-1)]. Both equations are estimated simultaneously. The important task is to determine the maximum lag order J. "

So my guess is that i should put these equations in the VAR in some way, but i have no clue how. Furthermore the paper says that i need to determine the maximum lag order "j" using the Schwarz criterion (SC). I dont know how to build this VAR in eviews5 or how i can interpritate the data?

Thanks in advance for any hulp you can give me guys and please try to explain it very simple to me because remember i know very little about eviews...

[donihue]

Typically in a VAR, you would use the stationary versions of all 3 variables (CR, Fans and H-index). This is easily done by clicking "Quick==>Estimate VAR" and filling in the variable names, using an arbitrary number of lags (default is 2). After that, in the VAR itself, click "View==>Lag Structure==>Lag Length Criteria" to apply the tests (which include SC), select the best lag length, then re-estimate the VAR using this number of lags. Et voilà!

Regards
Donihue

[Jeanvion]

First of all, thanks for the answer! I included the parameters in
the appendix, is this the correct methode?

[donihue]

No! You should put "d_fan d_std" (and whatever your third variable is) under "Endogenous" and leave only "c" under "Exogenous"
Regards
Donihue