City Research Online

Abstract

Credit spreads are important financial tools, since they are used as indicators of economic progression, investment decisions, trading and hedging, as well as pricing credit derivatives. Their role has become more significant for the European fixed income markets since the introduction of the Euro, which reshaped the mechanics of the financial environment. The introduction of single currency provided the means for a pan-European economic growth and cross-border development, liberalized a vast inflow of capital which was once fragmented into different currencies, and provided the dynamics of cross-border investments around a unified legislative framework. Thus, the main subject of the thesis is to provide further insight into and investigate the nature and the dynamics of credit spreads of European corporate bond indices during the credit crisis period.

Traditional quantitative credit risk models assume that changes in spreads are normally distributed but empirical evidence shows that they are likely to be skewed and fat-tailed, and if they are ignored then the calculation of loss probabilities will be seriously compromised. Therefore, the first area of investigation aims to provide further insight into the dynamics of higher moments and regime shifts in credit spread changes by applying a GARCH-type model that allows for time-varying volatility, skewness and kurtosis, as well as a Markov regime-switching GARCH specification to capture the structural changes in the volatility of credit spreads. Furthermore, a comparison of the different specifications is undertaken in order to assess which model better fits the empirical distribution of the data and produces best Value-at-Risk estimates. The results presented have significant implications for risk management, as well as in the pricing of credit derivatives.

The second area of investigation is to assess and evaluate time-varying correlation of credit spreads. Different multivariate GARCH models, such as Orthogonal-GARCH, the Constant and Dynamic Correlation GARCH models, Risk Metrics and Diagonal-BEKK, are applied to examine the behaviour and dynamics of time-varying correlation. Additionally, the performance of the proposed models is examined by determining whether they produce accurate VaR estimates. The study finds evidence in support of time-varying correlation coefficients between credit spreads which appears to be market dependent and has implications for pricing of derivatives, portfolio selection, trading and hedging activities, as well as risk management.

Finally, the impact of economic determinants of credit spreads such as the risk-free rate, inflation, as well as equity and commodity indices and volatilities, are investigated over different market conditions using regime switching models. The results highlight how the effect of the determinants on credit spreads varies across different market conditions and point to the existence of non-linear relationship between the determinants and credit spread changes. The study reveals that the regime dependent determinants have significant explanatory power only in the high volatility regime. Finally, it is shown that the feed-forward neural network model out-performs the other specifications applied in this study in terms of estimating out-of-sample mean forecasts.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HG Finance
Departments:	Bayes Business School > Finance Doctoral Theses Bayes Business School > Bayes Business School Doctoral Theses

Export

Downloads