Text

theses

In today’s globalized knowledge economy, technological knowledge plays an increasingly important role. Nowadays, cities and clusters cannot rely exclusively on local knowledge sources, but they need to combine local with complementary geographically distant (trans-local) knowledge sources. This dissertation contributes to the literature on the changing geographic composition of knowledge connections, and the complementarity of distant and local connections. We do this by providing a more detailed picture of how the spatial distribution of these connections is changing, and how they interact with one another across a mix of developed and developing country cities. In particular, we look at 62 cities to see how the geographic structure of their knowledge sourcing has been changing, both at the level of city dyads and in the overall structure of the worldwide knowledge network between cities. Using US patent citation data for patents invented in these 62 cities worldwide, our first study explores the nature of the association between local, trans-local and international citations. Our results show that in all cities there is a significant association between international and local citations, and that an increase in international citations leads to an increase in local connections. We also find that this effect is accentuated in highly innovative cities when compared to relatively lower innovative cities in our dataset.
Our second study looks at dyadic relationships for all possible city pairs in our city dataset, and examines the determinants of the level of knowledge outflows and knowledge inflows between them. Our results show that knowledge sourcing patterns between individual cities have varied with the extent of their co-specialization of activities, their relative position in the international knowledge network and their degree of engagement with general purpose technologies.
Using social network analysis techniques, we construct a unidirectional network of cities in our third study, since backward citations point in just one direction to prior knowledge sources. We observe how this network changes during our time period both in the aggregate and at the level of five selected sectors. The nodes in our network represent cities while the edges represent citations from one city to another. We calculate network statistics such as degree strength and eigenvector centrality to determine which cities have gained influence over time and which cities have become relatively less important. We find some developing cities have gained substantial influence over time especially in the network of patents in the ICT and other electrical equipment technological fields.

ETD_9221

Ph.D.

Includes bibliographical references

by Salma Zaman

doi:10.7282/t3-ma7a-7646

ETD doctoral

The author owns the copyright to this work.