درباره الگوها و عوامل موثر بر گسترش جهانی فن آوری های نوین

Taking a largely empirical approach this paper addresses the global spread of new technologies by defining two diffusion margins – the extensive, referring to the spreading of first use across economies and the intensive, referring to the intensity of use within economies. Using data relating to mail services we indicate the relative importance of the intensive and extensive margins in global diffusion over time. Using data on steamships and the basic oxygen process for steelmaking we also explore whether there are international spillovers in the diffusion process. We find evidence of spillovers which appear more likely to be negative than positive.

The benefits of new product and process technologies only arise as those technologies are diffused i.e. spread across their potential markets. In many cases these markets are global and technology eventually encompasses the whole world. However, as Stoneman and Battisti (2010) argue, the literature on the diffusion of new technology, with some exceptions of course (e.g. Grossman and Helpman, 1993 and Perkins and Neumayer, 2005), tends to take little note of patterns and determinants beyond national boundaries. Although there is a considerable literature on diffusion at the industry level, the firm level, and the national level, most of this literature implicitly appears to ignore events outside the domestic economy. Even the literature that undertakes international comparisons of diffusion in different countries largely treats each country as a freestanding unit rather than part of a global economy. Taking a largely empirical approach this paper addresses the global spread of new technologies by defining two diffusion margins – the extensive, referring to the spreading of first use across economies and the intensive, referring to the intensity of use within economies. The empirical work is based upon data in the Historical Cross Country Technology Adoption Data Set (HCCTAD) collected by Comin and Hobijn, 2003 and Comin and Hobijn, 2004.2 We first explore changes in the two defined margins over time and their relative contribution to changes in the extent of worldwide use over the diffusion time profile. Using the history of mail services (1830–1990) as an example, we find that diffusion at the extensive margin is complete long before diffusion at intensive margins and the relative contributions of changes in extensive and intensive margins to worldwide diffusion change over the diffusion process. We next address how diffusion in one country (the intensive margin) may be affected by changes in use in other economies i.e. whether there are international spillovers. We argue that the larger part of the appropriate diffusion literature takes the view that diffusion patterns are the result of the adoption decisions of locally based production units and emphasises two main approaches: (i) epidemic hypotheses relating to information spreading and risk reduction; and (ii) decision theoretic arguments based on rank and/or stock effects impacting upon the profitability of adoption. We consider the simple extension of both approaches to allow for inter-country spillover effects. In the former approach however the spillover effects will be positive, in the latter probably negative. The HCCTAD provides good data on the diffusion of particular technologies in different countries that enable an empirical approach to the spillover issue. The empirical work herein considers two technologies that have partly been chosen on the grounds of data availability but also selected because there is a history of their prior study. The first study addresses the switch from sail to steam (see Harley, 1971) over the period from 1809–1938.3 To this example we apply an epidemic model of diffusion. We find some limited evidence in favour of positive international spillovers. The second technology is the basic oxygen process in steel making (see Oster, 1982) where the data coverage is from 1952–1992. With this example we are able to illustrate both rank effects and most importantly some negative international spillover effects. This suggests that as diffusion proceeds and usage extends more widely in the world so the returns to later adopters are reduced and this delays their date of adoption. The paper closes by drawing conclusions.