X[t-1] | 0 | 3 | - | - |
X[t] | 3 | 5 | 3 | - |
Y[t-1] | 6 | 5 | - | - |
Y[t] | 4 | 0 | - | 5 |
Z[t-1] | 6 | 5 | - | - |
Z[t] | 8 | 9 | - | - |
So the difference is only taken for changes to and from zero i.e. D.Var_A or D.Var_B.
For changes between values other than zero, it doesn't really matter if they are represented by a - or a 0.
Does this make sense? I always find it difficult to articulate a problem like this.
What's the best way to go about this?
0 Response to Panel Data Conditional Difference
Post a Comment