Di Porto, EdoardoRevelli, Federico2017-10-242017-10-242009https://hdl.handle.net/2445/116984This paper explores for the first time the consequences of centrally imposed local tax limitations on the modelling and estimation of spatial auto-correlation in local fiscal policies, and compares three spatial interaction estimators: a) the conventional maximum likelihood estimator that ignores censoring; b) a spatial Tobit estimator; c) a discrete hazard estimator. Implementation of the above empirical approaches on the case of local vehicle taxation in Italy provides a reasonably coherent picture in terms of the direction and size of the spatial interaction process, and offers a plausible spatial interpretation of the race to the top in provincial vehicle taxes.36 p.application/pdfengcc-by-nc-nd, (c) Di Porto et al., 2009http://creativecommons.org/licenses/by-nc-nd/3.0/es/Política fiscalImpostos sobre automòbilsCorrelació (Estadística)Fiscal policyAutomobiles taxationCorrelation (Statistics)info:eu-repo/semantics/workingPaperinfo:eu-repo/semantics/openAccess