Essays on credit spreads

This thesis focuses on an empirical analysis of credit spreads from three different perspectives. The first analysis investigates the price discovery process of credit risk by means of single-name credit spreads obtained from four main markets, namely the credit default swap (CDS), the bond, the equity and the option markets. This study is the first which quantifies the credit risk price discovery by considering the four markets jointly. By using two main measures of price discovery derived from widely used common factor models, results show that, during periods of high volatility, price discovery occurs primarily in the option market, whereas, during stable periods, it takes place in the equity market. Furthermore, evidence of volatility spillovers from the option market towards the other markets is found in highly volatile periods. Finally. an innovative way to use GARCH volatility models is introduced, which allows to generate time-varying measures of price discovery. The second analysis examines novel trading strategies based on the information about the price discovery of credit spreads obtained from the CDS and equity markets for US and European single-name obligors. The profitability of the new strategies is compared with that of the popular "capital structure arbitrage (CSA)" strategy. The latter is based on the discrepancy between the CDS spread and its equity-implied spread but does not take into account the price discovery process in credit spreads. Results show that the new strategies outperform the CSA strategy during the financial crisis of 2007-2009. Moreover, the correlation of the new trading rules with the CSA is low or negative during the crisis, which suggests that by combining new and old rules, portfolio managers could reap diversification benefits at times when diversification, and hence risk reduction, was hardly achievable. The final contribution adds to the literature on credit spreads by analysing, for the first time, the forecasting performance of both linear and nonlinear models of CDS spreads. The statistical and economic significance of the models' forecasts are evaluated by employing various metrics and trading strategies, respectively. Linear models, in particular, generate positive Sharpe ratios for some of the strategies implemented, thus placing some doubts on the efficiency of the European CDS index market.EThOS - Electronic Theses Online ServiceGBUnited Kingdo