Do Trade Partners Matter? Estimating the Impact of Peer Effects on Sectoral Innovation Within Global Value Chains

Abstract

This paper is the first to employ social network analysis and peer-effects theory to investigate how knowledge spillovers occur through Global Value Chains. With data from the World Input-Output Database (WIOD) and the US Patent and Trademark Office (USPTO), I explore three main questions. First, can network location and centrality predict sectoral innovation outcomes? Second, does the inventive behavior of direct trade partners have a causal impact on a sector's own tendency to patent? Finally, do "offshoring relationships," characterized by high-skilled sectors importing inputs from low-skilled sectors, hurt innovation outcomes for less-skilled sectors and developing countries? The results of this study show that measures of network location and centrality are positively and statistically correlated with growth in invention, even after controlling for standard covariates from the literature. I propose the use of "peers-of-peers" instrumental variables which prove to be a relevant instrument in the context of the WIOD Data. I find that the innovation of trade partners to which a country-sector sends inputs does not have a causal impact on the focal country-sector's innovation outcomes. However, It appears that in a subset of the network characterized by only offshoring trade relationships, inventive behaviors of direct trade partners have a causal impact on innovation outcomes of the focal sector. Furthermore, the instrumental variable exercise has an added advantage in that it allows for better differentiation between links that are offshoring and otherwise. Although initial tests of offshoring through link-deletion exercises were inconclusive, the instrumental variable exercise yields significant results in favor of the offshoring effect. The magnitude of the impact of peer effects is smaller for the offshoring subset of links than in the subset of "non-offshoring" links characterized by high-skilled sectors sending inputs to low-skilled sectors. The findings and the methodology employed in this study have important implications for academics and policymakers concerned about the development prospects of GVC participation for less-skilled and less-developed countries (LDCs).