This thesis focuses on the empirical investigation of Credit Default Swap (CDS) spreads and return dynamics for listed corporates in the US, UK and EU. Academic interest in CDS market is continuously growing and this thesis aims to provide a better understanding of the CDS market dynamics. Specifically, this thesis explores three critical areas of research interest for the CDS market with each Chapter Two, Three and Four focussing on a specific aim, objectives and research questions within the context of the overall thesis. The thesis is largely based on three separate but broadly related research studies. The first study, Chapter Two, explores the dynamics of quarterly CDS spreads for corporates in US, UK and EU for the three major economic conditions namely; pre-crisis, crisis and post-crisis period. This study is the first to explore such a wider sample domain both in terms of the geographical coverage as well as the period of analysis. CDS spreads are regressed using both accounting based ad-hoc measures as well as theory driven market based variables, individually as well as collectively in a single combined model. This study documents the changing nature of spread predictor variables based on the sub-period of analysis and find the market based variables to be more closely aligned to spreads than their accounting counterparts. This study proposes the use of both information sets as additive rather than substitutive within the CDS pricing framework. This study also tests the effect of bond market liquidity dynamics and CDS market liquidity effect on CDS spreads and finds spreads in the post-crisis period to be plagued by both bond market and CDS liquidity dynamics. This study concludes that CDS spreads in the post-crisis period may be plagued by non-default driven factors and should not be considered as pure measure of corporate credit risk. Thus signals from CDS market should be carefully considered in conjunction with other financial market indicators before drawing policy implications. The second study, Chapter Three, evaluates the effect of the interest rate, quantitative easing and fiscal policy announcements in US and UK on corporate CDS returns. The unprecedented interventions announced by government and Central banks to contain the effect of the financial crisis provides the motivation for this study. This study measures the effect of these announcements on corporate credit risk by estimating daily CDS returns which is a better time series measure of corporate credit risk than CDS spreads or equity returns as used in past studies. This study notes an opposite effect of interest rate announcement where credit risk for firms following the interst rate announcement decreased for US corporate while it increased for UK corporates. Across both US and UK, corporate credit risk tends to be lower following QE announcements; highlighting its popularity during the financial crisis. Fiscal policy announcements are characterised by minor improvement in corporate credit risk which is short lived. By comparing pre and post announcement days abnormal return, this study finds that median abnormal return following US policy interventions were higher in post announcement days in US while an opposite effect can be noted for the UK corporates. This study concludes that policy interventions in US were more effective in stablising corporate credit risk for US corporate which policy announcements in UK were not effective. This study also tests the differential effect following policy interventions across corporates sampled based on sector, credit quality, frim size and CDS liquidity. No other study have undertaken such a detailed sub-sample analysis across policy announcements in US and UK and the findings underline the theme that firm specific heterogeneity leads to differential effect of policy announcement on corporate credit risk. The third study, Chapter Four, attempts to provide evidence of the generalizability of the Fama and French (FF) asset pricing model to the CDS market. The test on generalizability of the FF model to the CDS market has not been attempted before and this study is the first to check the external validity of the FF model with an aim to test if the model works for the CDS market. The findings from the portfolios returns indicate the average daily excess returns are not perfectly aligned as expected to the book-t-market, operating profitibility and investment factors and expose variations in average return sufficient to provide strong challenges in asset pricing tests. The relationship between the portfolio type and average excess return trend is also found to fluctuate based on the sub-period of analysis. Apart from testing the external validity of the FF model, this study also aims to access the external validity of the default risk hypothesis, by testing if the default risj is proced in the cross section of CDS returns and if the FF factors; SMB and HML factors are proxying for default risk in the CDS returns. The finding indicates that it is unlikely that SMB and HML are proxying for default risk. Overall, the findings from this study indicates the FF three factor (3F) and FF five factor (5F) model can be generalised to the CDS market, between the two models the 5F model is a better asset pricing model for the CDS market. This study goes a step further and queries if the FF factor model for the CDS market can be improved on by augmenting it with a default driven factor. Augmenting both the 3F and 5F model with default factor results in at best a marginal improvement to the models' explanatory power across the sub-periods analysed in this study. Hence for reasons of parsimony, this study suggest the FF 5F model to be preferred asset pricing model for the CDS market. Notwithstanding these separate contributions, overall this thesis contributes to a better understanding of CDS spreads, CDS returns and thus the CDS market in general. The past decade have seen a wealth of literature focussing on CDS market and the knowledge and understanding of the CDS market dynamics is being continuously refined and expanded. The findings of this thesis will provide useful insight and a deeper understanding for a variety of stakeholders including regulators, market participants, the financial community and the academic community at large to be able to better understand an important source of credit risk information.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:695850 |
| Date | January 2015 |
| Creators | Pereira, John |
| Contributors | Sorwar, Ghulam ; Nurullah, Mohamed |
| Publisher | Kingston University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://eprints.kingston.ac.uk/35845/ |
Page generated in 0.0026 seconds