Ricerc@Sapienza

In this paper, techniques proper to complex networks studies have been applied to analyze features of the investment styles and similarities in the Italian pension funds. The analysis has been developed through interdisciplinary approaches. First, we look at the node degree distributions; next, we consider the centrality measures, like betweenness and closeness. Results indicate that the network of funds is dense and assortative, with short path lengths. Moreover, through community detection algorithms, it is found that many funds show similar features. In particular, the network of benchmarks is far from being dense, is characterized by hubs, and is disassortative. Furthermore, the insertion of weights does not produce dramatic changes in the centrality measures, but it blurs the communities. Still, the k-core and the highest k-shell do properly evidence the most popular benchmarks. In conclusion, the network structure of the Italian pension funds, without taking into account information from weights, seems to contain already sufficient information for detecting similarities in investments styles.