Economic diversification as a percolation process

Abstract

In this thesis, we will use network theory to model some of the mechanisms that influence the economic development of countries. In the Economic Complexity framework of Hausmann and Hidalgo (2011), the production of any product requires a combination of specific capabilities, representing different skills and other non-tradable inputs. The relatedness between products in terms of their capability requirements can then be inferred from country-product export data (Hidalgo and Hausmann, 2009, Caldarelli et al., 2012). This leads to a network representation of the economy called the product space (Hidalgo et al., 2007). In this network, products are connected if they are produced by the same (type of) countries, suggesting they require similar capabilities for their production. It has been shown empirically that countries tend to develop products that are 'nearby' in the product space, building upon capabilities that are already present. This leads to the interesting perspective of viewing economical development as a diffusion process on a complex network. Here we take a theoretical approach and instead of starting from the data, try to capture the mechanisms that govern this diffusion process in a simple theoretical model that describes how the accumulation of capabilities can lead to economic growth. We construct a network model of the product space as the one-mode projection of a randomly generated bipartite country-capability network that indicates the capability requirements of every product. We then describe economic development as a percolation process on this product network. We also incorporate in the model the effects of knowledge spillovers, in which exchange of capabilities between industries enables economic growth. The model can explain the sudden 'take-off' of some countries' economies and stagnant development of others, and provides insight in how the distribution of capabilities can affect economic growth.