does not adequately account for the relationship
between the dependent and the independent variables when the whole sample period
is examined. As a result of a structural break, the coefficients β1 and β2 have changed,
and the model for the whole sample period is misspecified. The explanatory power of
two regressions run separately for the pre-IFRS and IFRS period is higher than that of
a regression run for the whole sample period.