This paper adjusts the non spatial model by Parry and Bento (2001) in a monocentric city structure. We compute the efficiency gains of various second best policies (defined as combinations of road toll schemes and revenue recycling programs) in the long run, i.e. when household relocation is possible. The results suggest that the cordon toll welfare gains exceed those from a flat kilometer tax, and might account for a very large part of the gains from a more intelligent pricing scheme, like a varying kilometer tax. However, these gains are conditional to specific types of revenue recycling, such as labor tax cuts or public transport subsidies. Further sensitivity analysis shows that the optimal type of revenue recycling depends on the level of inefficiency in the provision of public transit prior to the introduction of the policy.