Contagion in Financial Networks
The recent financial crisis was caused, in part, by the spread of “contagion” through a highly-connected network of financial institutions. Network models can be used to understand how loan defaults cascade through the balance sheets of banks that have contractual obligations to each other, and whether increasing the number of connections in the graph amplifies or dissipates the effects of liquidity shocks. They can also be used to predict the global effects of catastrophic events, such as a large institution filing for bankruptcy, and simulate stress tests of the financial system. Glasserman and Young have a working paper for the Office of Financial Research which gives an overview of important questions about the relationship between network structure and financial stability, and examines several models of the spread of contagion. They find that leverage plays a key role in determining the effect of increasing connectivity on stability, and briefly discuss “information contagion”, which is triggered by changes in market perception and can lead to general crises of confidence. Another key observation they make is that opacity (i.e., incomplete information about the network) may directly contribute to the spread of contagion, which implies that regulators should focus efforts on making the network of obligations between institutions more transparent.