October 2007 (updated March 2008)
ABSTRACT: I propose that the diffusion of a new technology within a country is influenced by the use of that technology elsewhere. I extend the classic epidemic (Bass) diffusion model by adding a world diffusion effect, where world diffusion is a simple average of diffusion across countries. Evidence to support the spillover hypothesis is found using country-level data on steam- and motor ship diffusion in 15 countries. The sign of the effect is unclear from theory and indeed I find both positive and negative estimates. The relationship between domestic and world diffusion does not appear to correspond to any simple leader-laggard model. Interpretation of the results is complicated by the fact that the model variables are non-stationary. Market segmentation (i.e. the relative importance of short- vs. long-haul traffic) could also explain some of the findings however I cannot control for this using the present data.
This is a joint paper with my supervisor Paul Stoneman.
ABSTRACT: It is argued that international diffusion involves two margins, the extensive and intensive respectively, reflecting usage extending to previously non-using countries and to increasing usage in countries post first use. We show that the extensive margin only plays a major role in overall international diffusion in the early years of the diffusion process. In the later years it is the intensive margin that is important. The issue that is raised from this for future research is how the internal and external margins are linked. In particular: (i) is intra-country diffusion affected by inter-country diffusion or the extensive margi? - a questions never asked, as far as we are aware, in the extensive body of domestic diffusion studies; and (ii) is inter-country diffusion affected by intra-country diffusion or the intensive margin?