Spillover across Eurozone credit market sectors and determinants

ABSTRACT

We examine spillover and its determinants among Eurozone sector level credit markets using time and frequency domain spillover approaches. Based on network theory and connectedness analysis, we identify the sectors that are major transmitters and receivers of spillover during normal and crisis periods. The rolling window analysis shows that short-run spillover among credit market sectors intensifies during global and Eurozone crisis periods. Further, using Bayesian model averaging, we find that overall financial conditions and stock market volatility are the main drivers of total and sector-level spillover. Our findings have important implications for policymakers and investors interested in Euro-area credit risk at the sector level.

KEYWORDS:

• Credit default swaps
• time and frequency domain spillovers
• network connectedness
• Bayesian model averaging
• crisis periods

JEL CLASSIFICATION:

• C51
• C52
• G11

Disclosure statement

The authors have no conflicts of interest to declare.

Notes

¹ Previous literature shows that country sovereign risk affects credit risk in the corporate sectors, especially banks and financial companies (see, among others, Kallestrup, Lando, and Murgoci 2016).

² For example, during the GFC, many US financial institutions sold protection on their mortgage securities and could not fully cover the defaults.

³ Providing an opportunity to predict the behaviour of a specific credit sector by using information from another credit sector.

⁴ Kizys, Paltalidis, and Vergos (2016) examine the regime-dependent relation between bank and sovereign CDS spreads in the Euro-area. Importantly, they reveal how the CDS market becomes ‘an important chain in the propagation of shocks through the entire financial system’.

⁵ Several recent studies consider the heterogeneity of market participants in their empirical analyses (e.g. Bouri et al. 2017b).

⁶ We use the terms ‘interdependence’ and ‘spillover’ interchangeably because the literature remains inconclusive in regard to a precise definition of interdependence/spillover or even contagion. Broadly speaking and in the context of sector-based analysis, spillover refers to the cross-sector transmission of shocks or general cross-sector spillover effects. A related, but more restrictive, definition refers to the term contagion, which reflects how cross-sector correlations increase during crisis periods relative to tranquil periods.

⁷ We use the information content in the CDS indices of 14 corporate sectors in the Eurozone. Our reliance on CDS spread as a proxy for credit risk is based on two main arguments. Firstly, since its inception in 1994, CDSs have constituted the largest portion of the credit derivatives, and CDSs have been used to transfer credit risk. Secondly, CDS represents an important measure of firm default risk, and is more suitable than bond spread, as shown by evidence from the literature that bond prices are often distorted by liquidity and tax problems (Fei, Fuertes, and Kalotychou 2017).

⁸ The reader can refer to Raftery and Painter (2005) for a more detailed discussion.

⁹ We use the BMA R package, available at http://cran.r-project.org/.

¹⁰ The use of change in CDS spread is common in the empirical literature (e.g. Bouri, Jalkh, and Roubaud 2017a; Bouri et al. 2018).

¹¹ As the spillover approach yields an N x N table of pair-wise spillover connections, we decompose the spillover table into symmetric and skew-symmetric components. A symmetric matrix with elements calculated as averages $\left(q_{ij}+q_{ji}\right)/2$, represents dissimilarities between objects, and, after transformation into a distance matrix, is visually represented using multidimensional scaling. A skew-symmetric matrix with elements $\left(q_{ij}-q_{ji}\right)/2$ represents the difference between transmitted and received spillover, embodying net transmission, and is shown using heatmaps.

¹² The financial conditions index Europe accounts for several macroeconomic and financial variables including measures of financial market stress, depressing economic activity, inflation, changes in perception and risk tolerance that alter risk premia, supply of credit, the success of policy measures, financial shocks, bank lending rates, money growth, spreads between government bond yields of different maturities, bank capital and liquidity, equity and securities issuances, bank and corporate bond yields, stock market returns of financial and non-financial institutions, volatility in equity and exchange rate markets, and correlations among financial variables, among others.