Systemic risk inherent in credit default swap indices and optimal execution strategies in limit order books for market makers

Abstract

This dissertation discusses two main subjects about the assessment of systemic risk inherent in credit derivatives and the optimal execution strategy in limit order books. In Chapter 1, we define systemic risk in two perspectives: the possibilities of simultaneous and contagious defaults. To explain these features, we employ a Marshall-Olkin copula model and an interacting intensity-based model. The model parameters are calibrated using credit default swap (CDS) index data, and the best time series models for each parameter are selected based on minimal prediction errors to forecast the dynamics of parameters. Chapter 2 investigates the evolution of systemic risk inherent in investment-grade (IG) and high-yield (HY) CDS portfolios and compares the portfolios before and after the global financial crisis. To quantify systemic risk, we propose a novel measure – the expected default rate (EDR), and implement the EDR under the one-factor copula framework with various dependence structures. In Chapter 3, we seek the market maker’s optimal trading strategy maximizing the profit and liquidating the inventory over a finite period. We develop the models of market and limit orders dynamics and derive the Hamilton-Jacobi-Bellman (HJB) equation for the optimal control problem. Then, we build an approximation scheme to find a solution of the HJB equation based on deep neural networks and compare the performance of the optimal execution strategy with that of other strategies by simulation.