Simulated ma(2) process but don't know how to estimate it

[smartbabe]

Hi there,

I have simulated an ma(2) process using the following code:

smpl @all
series e=nrnd
smpl @first+2 @last
series y=0.7-2*e(-1)+1.35*e(-2)+e

However, when I tried to estimate it with the following code:

y c ma(1) ma(2)

I was not able to recover the original regression coefficients.

Can anyone help me with this? I have spent a considerable amount of time looking for related information on the internet but I was not able to find anything related to this. Thank you very much!

[startz]

Those MA coefficients may not be invertible. Try it with different coefficients. And how large was your sample?

[smartbabe]

Those MA coefficients may not be invertible. Try it with different coefficients. And how large was your sample?

Is it that sometimes MA coefficients are not invertible? That's good to know.

Does that mean that my code is right?

When I tried to estimate it with the following code:

y c e(-1) e(-2)

I got the simulated coefficients back.

I tried with sample sizes of 100, 1000, and 10000.

[startz]

Your code is right.

Here's an example:

Code: Select all

series y = 2 + 2*e(-1) + e ls y c ma(1)

The estimated MA coefficient will be about 0.5. That's because for an MA(1) coefficients of theta and 1/theta imply the same MA process.

[smartbabe]

Your code is right.

Here's an example:

Code: Select all

series y = 2 + 2*e(-1) + e ls y c ma(1)

The estimated MA coefficient will be about 0.5. That's because for an MA(1) coefficients of theta and 1/theta imply the same MA process.

Thanks a lot! Just a quick side question: how do I create a command file in eviews like a do file in stata?

[startz]

File/New/Program

[smartbabe]

File/New/Program

Thanks. You are the best. =)