The different structure of economic ecosystems at the scales of companies and countries
A key element to understand complex systems is the relationship between the spatial scale of investigation and the structure of the interrelation among its elements. When it comes to economic systems, it is now well-known that the country-product bipartite network exhibits a nested structure, which is the foundation of different algorithms that have been used to scientifically investigate countries' development and forecast national economic growth. Changing the subject from countries to companies, a significantly different scenario emerges. Through the analysis of a unique dataset of Italian firms' exports and a worldwide dataset comprising countries' exports, here we find that, while a globally nested structure is observed at the country level, a local, in-block nested structure emerges at the level of firms. Remarkably, this in-block nestedness is statistically significant with respect to suitable null models and the algorithmic partitions of products into blocks have a high correspondence with exogenous product classifications. These findings lay a solid foundation for developing a scientific approach based on the physics of complex systems to the analysis of companies, which has been lacking until now.