J. Paul Dunne, Eftychia Nikolaidou, Ron P. Smith
November 12, 2005
Richardson's action-reaction model of an arms race has prompted a considerable body of research that has attempted to empirically estimate such models. In general these attempts have been unsuccessful. This paper provides an extensive reconsideration of the estimation issues and using some recent developments in time-series econometrics, provides a comprehensive analysis of data for Greece and Turkey. It finds evidence of some form of cointegration between the military expenditures in both countries, but not of Richardson arms race type.