This paper models for the first time a spatial process in local tax policies in the presence of centrally imposed fiscal limitations. Focusing on the frequently encountered case of a tax rate cap, we evaluate three empirical approaches to the analysis of spatially dependent limited tax policies: (i) a Bayesian spatial approach for censored dependent variables; (ii) a Tobit corner solution model augmented with a spatial lag; (iii) a spatial discrete hazard model. The evidence arising from an investigation of severely state‐constrained local vehicle taxes in Italy suggests that ignoring tax limitations can lead to substantial underestimation of inter‐jurisdictional fiscal interaction. Copyright © 2011 John Wiley & Sons, Ltd.