Liquidity and Resilience of Credit Network: a Simulation Study

We consider a model of credit network introduced in [1], and extend the analysis to account for properties empirically observed in real credit network, notably degree-preference transaction regime, risk-averse and external shocks. Trade-off between transaction path length and capacity of path is also studied. It is shown that credit network does not need to be very dense to obtain high liquidity level. Network size has no effect on the success probability for BA graphs. Credit network is robust, especially with respect to the condition of edges failure in ER graph. Transactions with degree-preference regime will perform better liquidity level than transactions with uniform regime at low network density. We find that risk of transactions will tend to become stable when average degree increases. Success probability is higher in the case of minimum-capacity transaction than it is for shortest-path transaction at low network density.