In this paper, we focus on the Granger casuality test in the presence of regime shift. We apply a vector autoregressive (4) model on Swedish series of industrial output and consumer price index for the period 1980:1 - 1998:6. To test for casuality, three different test methods namely the single equation Likelihood Ration test, the systemwise Rao's F-test and the Bootstrap test, have been used in this study. We show that when the assumption of parameter constancy is violated, due to the occurrence of the structural changes, Granger casuality tests can provide misleading inference about the underlining relationship of casuality.