Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. This version: July 8, 2013
Terms of use:
Documents in
AbstractIn 2006, a study, published in the Journal of Public Economics, employing a panel regression of 200 U.S. counties across 20 years, found a significant elasticity of homicides with respect to firearms ownership. Based on this finding the authors made the public policy recommendation of taxing gun ownership. However that study fell prey to the ratio fallacy, a trap known since 1896. All the "explanatory power" (goodness-of-fit-wise and significance-wise) of the original analysis was due to regional and intertemporal differences and population being explained by itself. When the ratio fallacy is accounted for, all authors' results can no longer be found. This is illustrated in this paper using a balanced panel from the data for 1980 to 2004. My findings are robust to (i) alternative specifications not subject to the ratio problem, (ii) using only data from 1980 to 1999 as in the original paper, (iii) using an unbalanced panel for 1980 to either 1999 or 2004, (iv) applying weighting as done by the original authors and (v) using data aggregated at the state level.JEL Classifications: C51; H21; I18; K42 Keywords: Gun Ownership; Social Costs; Ratio Fallacy; Spurious Regression a Thanks to participants of a Brown Bag Seminar held at Philipps-University Marburg and to participants of a Seminar on replication studies held at University of Göttingen both in which some good thoughts on the problem were had. Thanks to Bernd Hayo for contributing the 'moving averages' solution for the noisy proxy mentioned in section 7 and to Florian Neumeier for contributing the growth model used in section 6.4.