Dangerous liasons and hot customers for banks

Abstract Understanding the correlation between different customers’ loss of creditworthiness is crucial to credit risk analysis. This paper describes a novel method, based on weighted network model, in which set firms, customers same bank, represent nodes while their links and weights derive from total transaction amounts. We explore contagion mechanism deriving transmission difficulties one customer other clients bank so highlighting areas where higher. use real proprietary data provided by illustrate proposed approach.