6. Conclusion

View the most recent version.

I n this paper we began by documenting the higher degree of geographic concentration among technology adopters than in plants overall. Motivated by this observation, along with frequently mentioned benefits of knowledge spillovers in agglomerated regions, this paper has investigated the presence of knowledge spillovers in technology adoption by analysing the pattern of technology adoption across plants and, if it exists, the scope of the knowledge spillovers. We used various rich sources of information and surveys to carefully identify knowledge spillovers from prior adopters of technology to potential adopters of technology, looking separately at the effects of various other agglomeration externalities—exogenous effects, local amenities and plant heterogeneity. We did this by analysing the scope of knowledge spillovers along the various dimensions of sectoral, geographical and technological proximities.

The key finding of this paper is that a plant's probability of adopting a specific technology is facilitated by the presence of local prior adopters of the same technology in similar industries. By identifying the knowledge spillover effects along the interaction of technology × industry × region × time, this investigation has overcome the difficulties of identifying unobservable-knowledge spillover that may affect technology-adoption decisions. Furthermore, by showing that the greatest spillover effects are from prior adopters in the similar- process-yet-different-product market, it provides a very convincing story that the effects identified are highly unlikely to be a result of spurious correlation at any level. The distinct monotonically decaying pattern of knowledge spillovers in all three dimensions of functional, geographical and technological distance confirms the result. This paper finds that knowledge spillovers from prior adopters are bounded along the three dimensions: geographical proximity, production-process proximity and technological proximity.