Assessment of Influential Nodes in B&R Financial Networks

Abstract

This paper employs a novel centrality metric, DomiRank, to identify influential markets in the Belt and Road (B&R) financial system. Initially, we estimate the cross-predictability of stock indices from 53 B&R countries and construct dynamic and directed networks spanning from 2014 to 2023. Subsequently, influential nodes are identified applying the DomiRank metric, and the applicability and efficiency are compared against five traditional centrality metrics. Our research indicates that: (i) B&R financial networks exhibit a distinct time-varying characteristic in their network structure and risk profiles across different years, with major events such as the Chinese stock market crash, the COVID-19 pandemic, and the Russia-Ukraine conflict being reflected in these cross-predictability networks. (ii) The influence of B&R stock markets is associated with geographical location and the level of economic development. Asian stock markets (especially RCEP members) and developed markets in other regions play a critical role in B&R networks. (iii) Different centrality metrics show variations in ranking results due to evaluated aspects. The simulated attack approach validates the effectiveness of DomiRank centrality in assessment of nodal importance in B&R financial system. It is also observed that both DomiRank and degree centrality have superior performance compared to other centrality metrics.

Keywords: Cross-predictability, Financial networks, DomiRank algorithm, Influential nodes