Estimating demand functions for developing countries before and after the end of the Cold War, Dunne and Perlo-Freeman (2003) found little evidence of any change in the underlying relationship. One concern with their analysis was that the use of cross-section averages might have obscured important time series effects. This paper deals with this issue by analysing their data using static and dynamic panel data methods. This produces evidence of a change in relationship and suggests that the focus in the literature on cross-section analyses has indeed limited our understanding of important dynamic processes at work within countries.
Notes
An earlier version of this paper was presented to the Second CesA/IDN International Conference on Defence Economics and Security in Mediterranean and Sub-Saharan Africa , Lisbon, June 2002. We are grateful to the participants for comments. Dunne is grateful to the ESRC for support under grant R00239388.