 # Matrix Singular Error

Hi,

I am running a FMOLS and a DOLS estimator to for cointegration with structural breaks and I keep getting the feedback:

'Error Matrix singular'

Can anyone help me determine what I am doing wrong since the FMOLS is working well.

Thank you.

0

The error `G0048 Matrix singular` most often occurs when attempting to either invert a matrix or solve a system of linear equations with a coefficient matrix in which not all of the columns are linearly independent.

Here is a simple example which will cause the problem

``````//Create a 100x3 random matrix
x = rndn(100, 3);

//Add a fourth column to 'x' which is
//identical to the second column
x = x ~ x[.,2];

//Create a moment matrix
mm = x'x;

//Attempt to invert the singular matrix
mmi = inv(mm);``````

Sometimes this occurs because some of the variables in the model are very similar. Other times this occurs, because the data was loaded incorrectly. You should add a print statement after the line in which the error occurs to inspect the singular matrix, this may help reveal the problem. aptech

1,728

0

Thanks 0

The error `G0048 Matrix singular` most often occurs when attempting to either invert a matrix or solve a system of linear equations with a coefficient matrix in which not all of the columns are linearly independent.

Here is a simple example which will cause the problem

``````//Create a 100x3 random matrix
x = rndn(100, 3);

//Add a fourth column to 'x' which is
//identical to the second column
x = x ~ x[.,2];

//Create a moment matrix
mm = x'x;

//Attempt to invert the singular matrix
mmi = inv(mm);``````

Sometimes this occurs because some of the variables in the model are very similar. Other times this occurs, because the data was loaded incorrectly. You should add a print statement after the line in which the error occurs to inspect the singular matrix, this may help reveal the problem. aptech
1,728
0

Thanks ### Have a Specific Question?

Get a real answer from a real person

### Need Support?

Get help from our friendly experts.