restrictions in SVAR

[startz]

Take a look at the Help built into EViews.

[adarshad]

and also consider that there is simultaneity endogeneity
Can you guide in this regard

Thanks for making it clear and open

[adarshad]

Can you exactly piont which object?you suggest var object? can you exactly pin piont or piont to an example?

[startz]

http://www.eviews.com/help/helpintro.html#page/content%2Fsystem-System_Estimation.html%23

[adarshad]

Thankyou for your reply but this doesnot include a var object so I need to choose among these or I can also include VAR object as suggested by @EViews Gareth

[startz]

That’s just a system of equations. It is not a VAR at all.

[EViews Gareth]

Man walks into an Art shop and buys a pack of crayons.

Comes back the next day and says "I want to draw something green. Can you show me how to use the red crayon to draw green?".

Shop owner: "You can't draw green with a red crayon. Why don't you use the green crayon?"

"I want to use the red crayon to draw green".

[adarshad]

Man walks into an Art shop and buys a pack of crayons.

Comes back the next day and says "I want to draw something green. Can you show me how to use the red crayon to draw green?".

Shop owner: "You can't draw green with a red crayon. Why don't you use the green crayon?"

"I want to use the red crayon to draw green".

good comment but man is innocent and clueless trying to get help

Do you think that GMM object will be suitable for a relationship that has endogeneity and Long T

[adarshad]

I wanted to assess impact of sentiment related to firms on returns of firms but also want to see how each N size effect this relationship. I therefore want to draw a simultaneous system, which I thought will be VAR where I can see effect of every right hand variables on left hand entity so that I can analyze that whether they variation in size of each N effects returns
that's why I wanted to draw a VAR system
but problem is that I want to verify how each N contributes into variation of returns and whether their size effects that variation
I guess that can only be done by a VAR where we can first form a full system and then can analyze further by applying restrictions on certain relationships and study only relationships which I require

[adarshad]

my first system of var will include all x and all y as equations and variables
But I need to portray that returns of a firm(y1) are effected by a sentiment of a firm(x1) itself and sentiment of other firms(x2,x3,x4.....)...similarly return of 2nd firm(y2) is determined by sentiment of firm itself(x2) and other firms (x1,x3,x4.....) but not by other y's.so i want to exclude other y from relationships
Therefore my assumption was that i can create exclusion restriction for other Y and use as instrument to determine svar.
My theory also supports simultaneity which can be solved by this way
I hope now you undertstand it.

[EViews Gareth]

You’re describing a System.

[adarshad]

I mean in equation for y1, i put zero for y2 y3 y4 y5 in Bmatrix of restrictions , which mean the effect of coefficents of y2Y1 , Y3Y1, Y4Y1 , Y5Y1 will become zero with relationship not existing practically
Same I do for all other equations for all relationships I require to impose restrictions and develop restricted var from previously existing unrestricted var for Y1 Y2 Y3 Y4 Y5 X1 X2 X3 X4 X5.The desried relationships will be restricted to zero in restriction matrix

[adarshad]

FOR EQUATION Y1 , only X1Y1 X2Y1 X3Y1 X4Y1 X5Y1 will remain and
for equation Y2 , only X1Y2 X2Y2 X3Y2 X4Y2 X5Y2 will remain
i draw impluse response and vaiance decomposition to study my desired relationships

[startz]

You probably should listen to Gareth. He's a very smart chap.

Perhaps you should write out the equations you want to estimate in math rather than in words. Don't worry about EViews. Just write out the equations completely. A simple example will do.