Supervised by Flora Bellone and Paolo Zeppini
In this PhD thesis, I explore two key structural changes: the growth of international Input-Output (IO) linkages and the rise of green industries, while putting a special emphasis on emerging economies. I investigate how these structural changes are characterised by industrial configurations that create feedback loop effects. I employ network modeling techniques within the framework of autocatalytic networks introduced by Jain and Krishna (1998). To the best of my knowledge, this research is the first attempt to use the autocatalytic networks lens to scrutinize structural transformations within international IO networks. My investigation encompasses three case studies which will be the basis for the three chapters: (i) the emergence of Chinese industries as the center of the global production network; (ii) the changes of input linkages within and across post-Socialist nations; (iii) the co-evolutionary growth of green and non-green industries in emerging economies.
Co-written with Michela Chessa and Dominique Torre
International economics
This article uses a network approach to study the relationship between trade agreements and trade flows. For the first time in the literature, hypergraphs are used to capture the topology of trade agreements. This topology will be compared to that of the commercial flow network. For our analysis, we focused on a snapshot of data from 2017 (before CETA was implemented and before Brexit). A modularity analysis carried out on both trade agreements and trade flows shows an imperfect correspondence between the communities of countries found within the two networks. The results justify Brexit as a means of reconciling networks of trade flows and agreements, and the CETA agreement as a confirmation of Canada's membership in the trade flow cluster including the EU.
Co-written with Xing Yafei, Dieter Mitsche, Konstantin Avratchenkov, Dominique Torre and Michela Chessa
Proceedings of the 2018 IEEE 4th International Conference on Computer and Communications (ICCC).
In this paper, we analyze modularity-based algorithms to obtain clusters of economic datasets. We investigate whether there are correlations between trade data flows between different countries and economic agreements between countries. We apply different methods to the two datasets and compare the results obtained by maximum weight matching. Surprisingly, wenote that there is not a very strong correlation between trade flows and agreements between countries.
Co-written with Adria Rius
The recent emergence of product-level input-output datasets provides an unprecedented opportunity to model supply chains. This paper introduces the Hierarchical Inputs Network (HIN), a method to map the supply chain of any specific product at the 4- and 6-digit HS code level. The method builds on the AIPNET product-level input-output dataset (Fetzer et al., 2024) and addresses an important limitation in this dataset, namely that it dœs not distinguish between direct and indirect relationships between products. This distinction matters because supply chains are tiered systems, and understanding how countries engage with those supply chains means knowing not just what inputs feed into a final good, but at what stage they do so. The HIN resolves this with an algorithm that retains only direct input-output linkages and structures inputs hierarchically. The method is replicable for any product in AIPNET, and compatible with international trade data. We validate the HIN on the automotive supply chain and show its scalability to other products. We conclude by proposing avenues for further research based on the questions, old and new, that this method enables us to address.
Sole author
This paper examines how the European Union’s Carbon Border Adjustment Mechanism (CBAM) could reshape market access for African exporters and assesses whether alternative trade directions can mitigate emerging constraints. Focusing on CBAM-covered sectors, namely aluminium, iron and steel, cement, fertilisers, and hydrogen, the analysis first estimates the potential carbon cost faced by African producers by reconstructing CBAM allowances using official EU Emissions Trading System (ETS) benchmarks and CBAM default values, complemented by a CBAM-compliant novel Input-Output approach to measure carbon intensity of each country's upstream emissions of CBAM industries. The results highlight strong heterogeneity across sectors and countries. Aluminium, being the product that produces the highest amount of carbon emissions. While CBAM is likely to increase costs and reinforce existing asymmetries in EU-Africa trade, several Southern markets, including within Africa and parts of Asia, emerge as viable alternatives with lower regulatory barriers. The findings suggest that trade reorientation, combined with deeper regional integration under the AfCFTA and greater downstream value addition, can form part of a broader strategy to reduce vulnerability to carbon-related trade policies and support Africa’s structural transformation.
Co-written with Antonio Andreoni and Richard Kozul-Wright
This paper examines the impact of the United States' "Liberation Day Tariffs," announced on 2 April 2025, on African economies. Marking a decisive break from the post-Bretton Woods free-trade consensus, these unilateral tariffs aim to reduce trade deficits, reshore supply chains, and strengthen US bargaining power. Although the initial round of tariffs was temporarily replaced by a 10 per cent flat rate for most partners (with harsher terms for China), to allow for bilateral deals and by a moderated regime announced on 31st July, the policy has already triggered global supply chain disruptions and heightened trade tensions. The consequences are particularly severe for African countries, many of which rely on preferential access to US markets to industrialise. This paper develops a theoretical framework to analyse the direct, indirect, and cumulative effects of the new tariff regime and applies it to assess the likely impact on African industrialisation prospects. It also explores the erosion of African market access and discusses potential alternative export destinations, highlighting broader implications for Africa’s role in a rapidly shifting global trade landscape.
Sole author
Africa is at a critical juncture as the climate crisis and major geopolitical and economic shifts reshape global supply chains, presenting both risks and opportunities. The continent remains largely specialised in low-value natural resource exports, with limited manufacturing and marginal integration into global value chains (GVCs). Foreign control and restrictive trade regimes have hindered industrial upgrading. This paper reviews the structural barriers African countries face in GVCs—including trade liberalisation pressures and reduced policy space—and assesses how initiatives like the African Continental Free Trade Area (AfCFTA) could foster industrial development through regional integration and coordinated policies. It argues for a strategic repositioning of Africa in global production, centred on regional value chains, stronger domestic linkages, and African-led manufacturing.
Co-written with Alexandre Ruiz
In this article, we study the dynamic dependence between "green" and "non-green" capabilities. To this end, we analyze the autocatalytic structure of product interactions using data on bilateral trade flows. The novelty is to provide an assessment of how green capabilities are associated with industrial structure. Furthermore, we show how domestic green capabilities could predict subsequent economic growth in polluting industries.
Co-written with Flora Bellone and Paolo Zeppini
We study the emergence of China as a dominant player in the global economy using the innovative concept of the autocatalytic set (ACS) introduced by Jain and Krishna (1998). We start by building a World Input-Output Network (WION) from the second version of the World Input-Output database (WIOD), which covers the period 2000-2014. We then identify the ACSs in the WION and explore both their scaling properties and their changes over the period. Our analysis shows the evolution of Chinese industries from peripheral positions to central positions of autocatalytic structures and highlights the intertemporal complementarity of local and global circular causality mechanisms in the rise of the Chinese economy.
Sole author
We study the changes in input linkages in post-Socialist economies between 1990 and 2021. We analyze the differences in the evolution of input shares between post-Socialist countries that joined the European Union (CEEU countries) and those that remained in the Russian sphere of influence and joined the Commonwealth of Independent States (CIS). This is done by analyzing an Input-Output network based on the EORA database. From this chapter, we detect different patterns between both groups of countries. CEEU countries tend to partake more in international linkages, while CIS countries tend to be more selfreliant.
Sole author
We use network modeling and the autocatalytic sets framework of Jain and Krishna (1998) to explain how the structure of Input-Output linkages creates clusters in which the effect of price-reducing innovation accumulates. This phenomenon is studied globally by modeling a network based on the World Input-Output Database (WIOD), covering the period 2000-2014. We identify clusters based on the Autocatalytic Set methodology, and we consistently identify lower input prices among cluster members compared to non-members of the clusters.