Modeling Stock Survivability Resilience in Signed Temporal Networks: A Study from London Stock Exchange

This paper examines the dynamic evolution process in London stock exchange and attempts to model stock survivability resilience in the financial networks. A big historical dataset of UK companies from London stock exchange for 40 years (1976–2016) was collected and conceptualized into weighted, temporally evolving and signed networks using correlation coefficients. Based on the legal definition of corporate failure, stocks were categorized into Continuing, Failed and Normal groups. Accordingly, we conducted analysis on (1) The long-term evolution process of the entire population with statistical inference and visualization. (2) Multivariate logistic modeling of survivability resilience using short-term network measures, degree ratio (\(r_{i}\)), node degree (\(k_{i}\)), and node strength (\(s_{i}\)). The results show an exponential market growth but with a “fission-fusion” behavior in network topologies, which indicates dynamic and complex characteristics of its expansion. On the other hand, regression and modeling outcomes show that the survivability resilience is correlated with \(k_{i}\) and \(s_{i}\). Moreover, the analysis of deviance suggests that the survivability resilience could be described, by and large, as a function of \(k_{i}\) since it contributes the most significant difference. The study provides a novel alternative to look at the bankruptcy in the stock market and is potentially helpful for shareholders, decision- and policy-makers.