Abstract :
Tests of Ricardian equivalence based on the Euler condition can be written in several mathematically equivalent ways. This paper shows that these alternative solutions may give rise to different empirical results when liquidity constraints are ignored. This may be due to misspecification, since a crucial assumption of the model is violated, or due to the different parameter restrictions imposed by the different solutions. The evidence suggests that misspecification is the most likely explanation. Previous evidence based on restricted specifications should be interpreted with caution
