Equation System Estimation

[wangwz]

I am now testing a system of regression in Eviews and it generates several tables. The first table gives t-stat, sd, etc. as usual. However, I am not quite sure what the null hypothesis in this test.

[EViews Gareth]

Your question makes little sense.

[wangwz]

Your question makes little sense.

Hi, I attached a picture and my question is the t-stat (circled). I don't quite understand the null hypothesis for the t-stat.

[wangwz]

Your question makes little sense.

[EViews Gareth]

The coefficient is zero

[wangwz]

The coefficient is zero

so you mean the coefficients in the system are jointly zero?

[EViews Gareth]

No.

The t-statistic on the line that starts with C(1) is testing whether C(1) is zero.
The t-statistic on the line that starts with C(2) is testing whether C(2) is zero.
The t-statistic on the line that starts with C(3) is testing whether C(3) is zero.
The t-statistic on the line that starts with C(4) is testing whether C(4) is zero.

It is a complicated system.

[wangwz]

No.

The t-statistic on the line that starts with C(1) is testing whether C(1) is zero.
The t-statistic on the line that starts with C(2) is testing whether C(2) is zero.
The t-statistic on the line that starts with C(3) is testing whether C(3) is zero.
The t-statistic on the line that starts with C(4) is testing whether C(4) is zero.

It is a complicated system.

Thanks very much. In this sense, may I ask how I can test whether C(1)s are jointly zero in this system?

[startz]

View/Coefficient diagnostics/Wald

[wangwz]

View/Coefficient diagnostics/Wald

Thanks very much for your reply. May I ask what I should type in the Wald test? I want to test whether in the system of equations, c(1)s from all individual equations are jointly zero.

[startz]

c(1) only takes on one value, even if it appears in multiple equations.

[wangwz]

c(1) only takes on one value, even if it appears in multiple equations.

So you mean actually the coefficients are identical across all equations in this system?

[startz]

Yes. If you want different coefficients give them different numbers.

[wangwz]

Yes. If you want different coefficients give them different numbers.

I see. Thanks very much! :D