Page 28 - Contributed Paper Session (CPS) - Volume 1
P. 28

CPS658 Sagaren P.
                  there  are  deterministic  co-integrated  relationships  among  variables-
                  deterministic terms in the Var(p) model are not present in the VECM(p) form.
                  (2)  If  there  are  stochastic  co-integrated  relationships  in  the  Var(p)  model,
                  deterministic  terms  appear  as  the  VECM(p)  form  in  the  EC  term  or  as  an
                  independent term in the VECM(p) form.

                  4.  Data Analysis
                      Based on the results of the of the Augmented Dickey Fuller (ADF) unit root
                  test both series have a unit root and are first order difference stationary. The
                  results show that both series are I (1) processes. To test for co-integration,
                  Johansen's test was used. The maximum lag length was set to 7 quarters and
                  an  autoregressive  order  of  p=2  was  selected  based  on  the  minimum
                  information criterion.Both series were found to be co-integrated with rank=1.
                  The next step was to investigate the model specification for the 5 cases below.
                  Case 1: If there is no separate drift in the VECM(p) form then the model is
                  given by:
                               ′
                      Δ =   −1  +  Δ −1 +. . . + −1 Δ −+1  +   from (2)
                                                                    
                                       1
                        

                  Case 2:
                  Suppose there is no separate drift in the VECM(p) form, but a constant 
                                                                                            0
                  enters only via the error correction term.
                  Consider the K-dimensional Var(p) process where  is K×1 and  =  + 
                                                                                
                                                                                     
                                                                                          
                                                                    
                  Suppose   =   are fixed K-dimensional parameter vectors. Then
                                0
                                          ′
                                 ∆ =   −1  +  Δ −1 +. . . + Δ −+1  + 
                                                                
                                                                             
                                    
                                                   1
                                       ′
                                      =  ( −1  −  ) +  Δ −1 +. . . + Δ −+1  + 
                                                                    
                                                                                
                                                      1
                                                0
                                       ′
                                                 ′
                                  =    −1  −   +  Δ −1 +. . . + Δ −+1
                                                                      
                                                   0
                                                        1
                                      ′
                                 =   −1  +   Δ −1 +. . . + Δ −+1  + 
                                               0+ 1
                                                                
                                                                             
                  where  = −   is the restriction of the intercept.
                                  ′
                                    0
                          0
                  Case 3:
                  There is a separate drift   and no separate linear trend in the VECM(p) form
                                           0
                  the following model is used.
                                          ′
                                 ∆ =   −1  +  Δ −1 +. . . + Δ −+1  + 
                                                                             
                                                                
                                                   1
                                    
                  Case 4:
                  There is a separate drift and no separate linear trend in the VECM(p) form, but
                  a linear trend enters only via the error correction term.
                  If  =  +   is a linear trend, we have  =  − 
                                                                    1
                                                                
                          0
                               1
                     
                                                           
                                 =  −  −   and Δ = Δ − Δ( −  −1 )
                                           0
                                                         
                                                                      0
                                 
                                                1
                                                               
                                      

                                             ′
                  Therefore from (3) ∆ =   −1  +  Δ −1 +. . . + −1 Δ −+1  + 
                                                      1
                                       
                                                                                   
                  Substitute  =  −  ; Δ −1  = Δ −1  −  ; and  =  −  −  
                                                          1
                                                                                1
                             
                                                                      
                                                                 
                                                                           0
                                       1
                                  

                                                                      17 | I S I   W S C   2 0 1 9
   23   24   25   26   27   28   29   30   31   32   33