1. Consider a VAR model for n variables, written in vector error correction form: AY, NC.Y-1+&t. (1) C is an n x n matri

[answerhappygod]

1 Consider A Var Model For N Variables Written In Vector Error Correction Form Ay Nc Y 1 T 1 C Is An N X N Matri 1 (80.26 KiB) Viewed 33 times

1. Consider a VAR model for n variables, written in vector error correction form: AY, NC.Y-1+&t. (1) C is an n x n matrix and & is vector white noise. The ‘unit root' properties of the data in the VAR depend on the nature of the matrix C. There are three possibilities: 1) all the variables in Y, are stationary; 2) all the variables in Y, individually are unit root processes and there is no cointegration; 3) all the variables in Y, individually are unit root processes but they are cointegrated (at least one linear combination is stationary). Match each of these possibilities to the corresponding restriction on the matrix C. (a) C=0 (C is a matrix of zeroes) (b) C =-1 (I is the identity matrix) (c) C#0 but |C| = 0 (the determinant of C is zero, which means that C is not linearly independent and therefore non-invertible) =