Theory dictates that when there is one cointegration relationship, it's coefficient in a VECM must be statistically significant and negative to maintain the long-run equilibrium.
Nonetheless, when there are two cointegrating relationships (Coint 1 and 2), should both their coefficients, C(1) and C(2), be negative and statistically significant or only the first coefficient, C(1)? Thanks
a, b, c, d, e, and f are real numbers.
Coint 1 Coint 2
Var 1 1.00 0.00
Var 2 0.00 1.00
Var 3 a d
Var 4 b e
Constant c f
VECM:
∆Var1(t)
Coint 1 C(1)
Coint 2 C(2)
∆Var1(t-1) C(3)
... etc
