Working Paper: NBER ID: w12812
Authors: Ajay K. Agrawal; Avi Goldfarb
Abstract: We report evidence indicating that Bitnet adoption facilitated increased research collaboration between US universities. However, not all institutions benefited equally. Using panel data from seven top engineering journals, Bitnet connection records, and a variety of institution ranking data, we find that medium-ranked universities were the primary beneficiaries; they benefited largely by increasing their collaboration with top-ranked schools. Furthermore, we find that the magnitude of this effect was greatest for co-located pairs. These results suggest that the most salient effect of lowering communication costs may have been to facilitate gains from trade through the specialization of research tasks. Thus, the advent of Bitnet -- and likely subsequent versions, including the Internet -- seems to have increased the role of second-tier universities in the national innovation system as producers of new, high-quality knowledge.
Keywords: Research Collaboration; Communication Costs; University Innovation; Bitnet
JEL Codes: O33; R11; Z13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| bitnet adoption (L86) | collaboration rates (D26) |
| bitnet adoption (L86) | collaboration rates for medium-ranked universities (A14) |
| bitnet adoption (L86) | collaboration rates with top-ranked schools (A14) |
| bitnet adoption (L86) | specialization in research tasks (C90) |
| bitnet adoption (L86) | role of second-tier universities in national innovation system as producers of new high-quality knowledge (O36) |
| quality of collaborating institutions (I24) | returns to bitnet adoption (D85) |
| geographic distance (R12) | collaboration rates (D26) |
| colocated pairs (Y80) | increase in collaboration rates (O36) |