Bipartite Network Projections of Multi-Locational Corporations: Realising the Potential

In this commentary, I evaluate the potential of bipartite network projections of multilocational corporations in quantitative structural analyses of world/global cities. My starting point is that the Geographical Analysis paper by Pazitka, Wojcik, and Knight (2020) paper provides a number of useful critical observations in this regard, but their analysis is also incomplete and at points potentially misleading. I make three points. First, in this literature the notion of "construct validity" is vague because there is no such thing as a comprehensive world city network: we are dealing with abstractions of a complex and multilayered reality at best, and bipartite network projections are no more than generic tools that can help capturing the multiple globalizations of specific sets of cities. Second, there is no reason to assume that a spatial interaction-based benchmark model can be used to assess the "structural validity" of these bipartite network projections. And third, rather than replacing bipartite network projections of multilocational corporations with bipartite network projections of project ecologies, this literature could above all benefit from geographers no longer borrowing projections from the social networks literature. Instead, we should try and come up with bipartite network projections that more adequately capture the geographical nature of how multilocational corporations/projects operate/unfold.