This work extends the contagion model introduced by Nier et al. (2007) to inhomogeneous networks. We preserve the convenient description of a financial system by a sparsely parameterized random graph but add several relevant inhomogeneities, namely well-connected banks, financial institutions with disproportionately large interbank assets, and big banks focusing on wholesale and retail customers. These extensions significantly enhance the model's generality as they reflect inhomogeneities as found in reality with a potentially decisive impact on system stability. Whereas well-connected banks and big retail banks have only a surprisingly modest impact, we find a significantly enhanced contagion risk in networks containing institutions with disproportionately large interbank assets. Moreover, we show that these effects can be partly compensated by a suitable regulatory response which demands additional net worth buffers for banks with above average volume of interbank assets. The stabilising effect is most notably achieved by a pure redistribution of equity capital without increasing its total amount.
capital buffers contagion contagious defaults inhomogeneities network models financial system stability