A probabilistic method for reconstructing the Foreign Direct Investments network in search of ultimate host economies

The Ultimate Host Economies (UHEs) of a given country are defined as the ultimate destinations of Foreign Direct Investment (FDI) originating in that country. Bilateral FDI statistics struggle to identify them due to the non-negligible presence of conduit jurisdictions, which provide attractive intermediate destinations for pass-through investments due to favorable tax regimes. At the same time, determining UHEs is crucial for understanding the actual paths followed by FDI among increasingly interdependent economies. In this paper, we first reconstruct the global FDI network through mirroring and clustering techniques, starting from data collected by the International Monetary Fund. Then we provide a method for computing an (approximate) distribution of the UHEs of a country by using a probabilistic approach to this network, based on Markov chains. More specifically, we analyze the Italian case.