Page 31 - Contributed Paper Session (CPS) - Volume 1
P. 31
CPS658 Sagaren P.
H0: H1:
Rank=r Rank>r Eigenvalue Trace Pr > Drift in ECM Drift in Process
Trace
0 0 0,2259 8,028 0,4625 Constant Linear
1 1 0,0206 1
0,603 0,4370
3
Table 4 shows the results of the “Co-integration Rank test using trace”
where there is no trends in the Co-integration space and VECM has drift terms
. : Rank=0 cannot be rejected
0
Since the drift term is unrestricted, there are still linear trends in the data,
but no deterministic trends in any Co-integration relations. In this application
a non-zero unrestricted intercept and no trends in the co-integrating
relationship may be plausible.
Case 4: There is a separate drift in the VECM and a linear trend in the Co-
integration space.
The trend is restricted to lie in the Co-integration space, but the drift term is
unrestricted in the VECM.
This model may be used to test for trend stationarity.
Table 5
Co-integration Rank Test Using Trace Under Restriction
H0: H1: Drift in
Rank=r Rank>r Eigenvalu Trace Pr > ECM Drift in Process
Trace
0 0 e 22,3309 0,1286 Linear Linear
0,4219
1 1 0,1992 6,4410 0,4060
Table 6
Hypothesis Test of the Restriction
Restricted
Rank Eigenvalue Eigenvalue DF Chi-Square Pr > ChiSq
0 0,4135 0,4219 2 0,43 0,8058
1 0,1987 0,1992 1 0,02 0,9016
Clearly (from Table 5) H0: Rank=0 can be rejected at the 15%
significance level whilst,H1 cannot be rejected, there is evidence of Co-
integration with restriction on the trend.The model may not be
appropriate i.e given the p-values in Table 6.
Case 5: No restrictions on the trend and intercept in the VAR model.
20 | I S I W S C 2 0 1 9