AN INVESTIGATION OF RICARDIAN EQUIVALENCE IN A COMMON TRENDS MODEL

To empirically discriminate between Ricardian and non-Ricardian behavi our, we have to use a method that distinguishes between expected and u nexpected, as well as permanent and transitory changes in taxes and go vernment consumption. One method that provides these distinctions betw een different types of changes is the common trends framework used in this paper, which is a VAR model with cointegrating constraints. In th e empirical study of US data, there is some support for the Ricardian hypothesis, but there are also some deviations from its predictions. H owever, the differences do not lend support to a Keynesian view of bud get deficits, but rather to the view of expansionary fiscal contractio ns. This can be motivated theoretically by uncertainty about future ta xes in combination with prudent households.