Models of systemic risk in financial markets

This thesis studies systemic risk in financial markets and how it emerges through dynamical and structural amplification mechanisms. In part (1) I study the dynamics and control of Basel leverage cycles. For this I develop a simple model of a financial system consisting of leveraged banks and an unleveraged fundamentalist investor (fund). Banks trade a risky asset with the fund and rely on historical information to estimate their portfolio risk. This risk estimate determines the banks' leverage limit. I show that these simple ingredients can lead to endogenous, irregular oscillations, which I call Basel leverage cycles. I then proceed to evaluate alternative regulatory capital requirements based on their impact on endogenous risk. I find that in the microprudential limit, when the bank is small and exogenous volatility is high, the optimal policy is simply given by a Value-at-Risk constraint. However, when the bank is large, the optimal policy is constant leverage. In part (2) I study contagion in financial networks for two examples. First, I study how intra-institutional linkages can amplify financial contagion when financial institutions are active in multiple over-the-counter markets. In particular, spillover within a diversified financial institution allows for contagion from one over-the-counter market to another. Using recent methods for coupled networks I illustrate that under certain circumstances, the presence of intra-institutional spillover can lead to the amplification of small shocks to the extent that trading across all markets collapses abruptly. Finally, I develop a simple model of social learning in the context of a financial network. I study how banks' portfolio decisions can synchronize if banks rely both on outside information and information from their social network to compute the expected payoff of an investment opportunity. In the same model, I propose a simple boundedly rational decision mechanism for endogenous network formation based on the information content of a bank’s neighbors' decisions.