Robots and the regionalization of global value chains

In this paper, we study whether and to what extent exposure to industrial robots leads to the regionalization of global value chains (GVC) for a group of seven European countries and ten manufacturing sectors. We use country-industry-year data on GVC participation for the period 1995-2018 from the OECD-ICIO database and we merge it with industrial robot data from the IFR. To assess the non-spurious long-run relationship between robots and GVC dynamics, we adopt a panel cointegration approach and dynamic OLS regressions, while we assess the direction of causality using a panel vector error correction approach. Our results suggest that, on average, higher exposure to robotization Granger causes a higher GVC regionalization, which is more pronounced when the source of foreign value-added moves from Asian economies to Eastern Europe. We also find that sectoral heterogeneity matters, since a stronger robot-induced regionalization of GVCs tends to occur in more upstream sectors and with high labour intensity. • We study the long-run effect of robotization on GVC regionalization • We use a panel cointegration approach after merging OECD, EU-KLEMS, and IFR data • Robotization pushes EU countries to insource more intensively from their EU neighbours • Robotization pushes EU countries to insource less from distant countries (e.g., Asia) • Results are stronger in sectors with higher relative labour endowment and upstreamness