Empirical evidence shows that government spending crowds in private consumption, a keynesian phenomenon. The current, state of the art, new keynesian models based on optimizing households and firms are not able to predict such a result. In this paper, we critically analyse fiscal policy in these models using a graphical framework as well as a formal model. Extensions aimed at generating crowding in, like useful government spending or rule of thumb consumers, turn out to be inappropriate. We argue that introducing productivity enhancing government spending could potentially lead to crowding in.