Bai and Carrion-i-Silvestre (2009) - Interpretation of Results

Hello everyone. I have run the GAUSS "Panelbreak" code to implement a panel unit root test with structural breaks (Bai and Carrion-i-Silvestre 2009). I was hoping someone could tell me how to interpret the matrix of "Estimated break points" in my results:

1.0000000 17.000000 0.0000000
2.0000000 0.0000000 0.0000000
3.0000000 9.0000000 15.000000
4.0000000 0.0000000 0.0000000
5.0000000 11.000000 0.0000000
6.0000000 5.0000000 19.000000
7.0000000 7.0000000 17.000000
8.0000000 15.000000 0.0000000
9.0000000 8.0000000 0.0000000
10.000000 13.000000 19.000000

Many thanks in advance.

 

1 Answer

0

Hello,

The "Estimate Break Points:" matrix, which is printed to the screen when you run the "Panelbreak" code, provides estimates of the structural breaks for each individual in your panel.

The first column in the matrix identifies the individual in the panel. In other words, each row in the table represents the results for a unique individual in the panel.

In this case, I can tell that you have 10 individuals in your panel because there are 10 rows in your matrix.

The remaining columns represent the estimated optimal breakpoints, 1-m.

In this case, I can tell that you specified 2 as the maximum number of breaks because there are two additional columns (three total columns).

Let's look at a few specific examples within the table.

  • For the first individual, in row one, there is a structural break estimated at the 17th observation and no additional structural break is found in the data.
  • For the second individual, in row two, no structural breaks are found in the data.
  • For the sixth individual, in row 6, a structural break is found at the 13th observation and a second structural break is found at the 19th observation.

You must login to post answers.

Have a Specific Question?

Get a real answer from a real person

Need Support?

Get help from our friendly experts.

Try GAUSS for 14 days for FREE

See what GAUSS can do for your data

© Aptech Systems, Inc. All rights reserved.

Privacy Policy