International trade and Foreign Direct Investments (FDI) are two phenomena in continuous
evolution, that play a fundamental role in shaping countries economies and development.
In today's world economy these two phenomena are also intimately related
through the processes of the Global Value Chain (GVC): in the last decades,
transportation costs and globalization had led to a fragmentation of the production
process and the building of proper productive networks. Usually coordinated by
transnational corporations these networks have the purpose of optimizing the
production process by placing different production stages in countries where they
are more cost-effective and, by doing so, link many world countries through various
FDIs and trade flows. Globally this is a huge phenomenon: in 2010 for example, these
international production networks accounted for approximately two thirds of both exports and imports of goods for the United States and France
(cf. UNCTAD World Investment Report 2013).
In this paper we want to formalize this relations and study the phenomenon in a multiple-networks
perspective: using BACI (cf. http://www.cepii.fr/) and a recently published FDI dataset (cf. ECB
WP 1554, June 2013) we build a graph where each world country is a node and can have two
different type of weighted directed edges, one related to goods trade and one to FDIs.
We first study general topological properties of these two sub-graphs finding, among other
properties, that they both have low cluster coefficients (on average 0.20 for trade and 0.04 for FDI)
and they are both disassortative. Moreover, looking at their nodes' cumulative strength distribution,
we discover that they have the same topological structures: both the international trade and the FDI
networks present a power law strength distribution, with almost identical exponents.
The edges weights of the two networks present a positive correlation and, controlling for countries
properties like GDP per capita, population, contiguity, etc., using the Heckman selection model we
found a significant positive elasticity of 0.12 of trade relative to FDI. However, differentiating the
world trade in the three main macro-sectors, i.e. primary (raw materials), secondary
(manufacturing) and tertiary (services), we find that this positive relation is not constant: for
services, in particular, it reverts, indicating that trade and FDIs in that sector are substitutes.
An important distinction in Foreign Direct Investments is between vertical and horizontal: given
that their distinction is not really sharp and ultimately depends on the actual (and unknowable)
structure of the specific transnational corporation, we study the interaction of the relation between
trade and FDI with upstreamness (cf. Antràs, AER 2012): this measure point out, for each product
category, its collocation in the production chain, from raw materials to the final good. Our results
indicate that worldwide trade is stronger for goods downstream in the production chain and that the
presence of FDIs fosters this relation, intensifying trade more effectively for goods closer to the