2sls estimation

[Gr8Mate]

I'm a noobe running Eviews7, I'm used to SPSS & have run 2SLS with it. Searching the forum and the help file has only confused me

Testing firm performance and ownership structure, OLS is:
Perf c Size Lev Growth Own1 Own2 Own3 Own4 etc1 etc2 etc3

Tests show that Own3 Own4 are endogenous. How do I run this in Eviews? What should I put in the top and bottom boxes?

Many thanks for your help!!!

Mark

[startz]

Put

Code: Select all

Perf c Size Lev Growth Own1 Own2 Own3 Own4 etc1 etc2 etc3

in the top box and the list of series you wish to use as instruments in the bottom box.

[EViews Gareth]

In the top box put your list of variables:
Perf c Size Lev Growth Own1 Own2 Own3 Own4 etc1 etc2 etc3

In the second box, enter the exogenous variables (size lev growth own1 own2), plus whatever other instruments you will be using for own3 and own4.

[Gr8Mate]

Many thanks for your help

"other instruments you will be using for own3 and own4"

I don't have any other instruments, own3 and own4 is all I have. This is where I'm stuck Any suggestions?

[startz]

Yes, you're stuck. But that's how 2sls works. You'd be in the same situation in SPSS or any other program.

[Gr8Mate]

If I put top box
Perf c Size Lev Growth Own1 Own2 Own3 Own4 etc1

bottom box
c Size Lev Growth Own1 Own2 etc1 etc2 etc3

would that work?

[startz]

Depends what you mean by "work." You'll get an answer, but you've changed the specification of the equation you're estimating.

[EViews Gareth]

Subject to the instruments satisfying the order condition too.

[startz]

Subject to the instruments satisfying the order condition too.

Well,

Code: Select all

tsls Perf c Size Lev Growth Own1 Own2 Own3 Own4 etc1 @ c Size Lev Growth Own1 Own2 etc1 etc2 etc3

does satisfy the order condition. (Did you mean rank?)

[Gr8Mate]

Further
The other instruments in the bottom box for own3 & own4 - can they be the same indt variables from the list of indt variables in the top box?

[EViews Gareth]

Subject to the instruments satisfying the order condition too.

Well,

Code: Select all

tsls Perf c Size Lev Growth Own1 Own2 Own3 Own4 etc1 @ c Size Lev Growth Own1 Own2 etc1 etc2 etc3

does satisfy the order condition. (Did you means rank?)

I just meant that he did have to have etc2 and etc3 in the instrument list - i.e. if he is going to have own3 and own4 as endogenous, he does need at least two more instruments over those variables included in the regressor list.

If we assume that etc1, etc2 and etc3 are single variables, then yes, he is currently meeting the order condition.

[Gr8Mate]

thanks for your help!!! I'll give it a try today