Economic Complexity: exploration of ranking metrics and investigation of causality between Complexity and Gross Domestic Product per capita.

Several political economists (e.g. Ricardo, Smith) tried to explain the economic growth phenomenon without reaching a univocal conclusion. In recent years, a new discipline called Economic Complexity is emerging over classical economic theories, trying to give an answer to the fateful question “what is the secret of the wealth of nations?”. According to Economic Complexity theories, it is possible to estimate the industrial competitiveness of a country just looking at its export basket, namely the products the country is able to export. In this approach, the trade data regarding exports is interpreted as a bipartite network in which countries are connected to the product they export. In order to classify countries and products belonging to this bipartite network, different algorithms have been developed, mutually correlating the complexity of countries to the one of products. At present, the commonly used Economic Complexity metrics are the Method of Reflections (MR) (Hidalgo & Hausman, 2009) and the Fitness and Complexity algorithm (FC) (Tacchella, Cristelli, Caldarelli, Gabrielli, & Pietronero, 2012). These two metrics diverge among them in the maths and in the results they supply. For this reason, a team from Politecnico di Torino have recently designed a new metric, called GENeralized Economic comPlexitY index (GENEPY) (Sciarra, Chiarotti, Ridolfi, & Laio, 2020) that is able to reconcile the two-existing metrics and furnishing a unique complexity ranking of countries and products. The first aim of this thesis is to analyse the three mentioned EC metrics, visualising the main differences and similarities among them. Particular attention has been paid to the complexity of products, because most of the works on Economic Complexity focused on the reliability of the results looking mainly at countries’ competitiveness. Thus, through a detailed analysis of products' complexities and by exploring their evolution over time, we show that products have a crucial role on this complex system: the increase of a country's economic complexity translates into the presence of new complex products in its export basket; the presence of niche products in the network, especially if exported by developed countries, compromises the reliability of the results. The second aim is to investigate the causal relationship between the EC metrics and the most currently used economic competitiveness index, the Gross Domestic Product per capita (GDP). As first suggested by Sugihara et al. (Sugihara, et al., 2012), to carry this analysis we use the Convergent Cross Mapping technique. We show that a wide set of countries presents a strong causality correlation between GDP and EC metrics, suggesting the possibility to predict one from another. The prediction of GDP using only the EC algorithms represents the toughest challenge. However, our results on this subject show the possibility to consider Economic Complexity as a driving force for economic growth.