Asymmetric and nonlinear comovements of credit default swap and bond markets: evidence from an emerging market

ABSTRACT

This study examines the equilibrium and price discovery relationships between credit default swap (CDS) and bond markets through the symmetric and non-symmetric cointegration approaches, focusing on Turkish sovereign markets. While the findings, obtained using the threshold autoregressive (TAR) and momentum threshold autoregressive (MTAR) models, confirm the asymmetric and nonlinear cointegration relationship between CDS and bond markets described in a limited number of studies in the relevant literature, they also, in contrast to other relevant studies, reveal the importance of data frequency and structural breaks in price discovery research. In this context, the results show that the CDS market is the leader in price discovery in high-frequency data and after the first sub-period, and the bond market is the leader in price discovery in low-frequency data and before the first sub-period.

KEYWORDS:

• Credit Default Swaps
• sovereign bonds
• price discovery
• nonlinearity
• structural breaks
• emerging markets

JEL CLASSIFICATION:

• G10
• G12
• G32

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data that support the findings of this study are obtained from the Bloomberg database, which is a paid subscription.

Notes

1. Brazil, Bulgaria, Colombia, Mexico, The Philippines, Russia, and Venezuela.

2. Brazil, Chile, China, Columbia, Malaysia, Mexico, Panama, Peru, Philippine, Poland, Russia, South Africa, South Korea, Turkey, Ukraine, and Venezuela.

3. Argentina, Brazil, Chile, China, Colombia, Korea, Mexico, Malaysia, Panama, Peru, Philippines, Poland, Russia, Turkey, Venezuela, and South Africa.

4. Brazil, China, Colombia, Mexico, Philippines, Russia, Turkey, Uruguay, and Venezuela.

5. Argentina, Brazil, Bulgaria, Colombia, Mexico, Panama, Peru, Philippines, Russia, and Turkey.

6. Argentina, Brazil, Bulgaria, China, Colombia, Indonesia, Mexico, Panama, Peru, Philippines, Poland, Russia, South Africa, Turkey, Ukraine, and Venezuela.

7. France, Germany, Italy, Japan, the UK, and the US.

8. These factors include, among many others, the cheapest-to-deliver (CTD) option and difficulties in short-selling the cash bond (Blanco et al., 2005); the different responses of the two markets to changes in the credit quality of reference entities (Zhu, 2006); measures of company-specific credit risk and liquidity, and of market conditions (Trapp, 2009); latent liquidity (accessibility of a bond), liquidity of the CDS contract, and the riskiness of the CDS market (Nashikkar et al., 2011); collateral quality, bond illiquidity, counterparty risk, and funding liquidity risk (Bai & Collin-Dufresne (2019)); contractual characteristics of CDS and bond contracts, technical aspects related to CDS and bond pricing and market forces affecting the bonds, the CDSs, or the basis directly (Augustin & Schnitzler, 2020).

9. Argentina, Brazil, Bulgaria, Chile, China, Colombia, Cote d’Ivoire, Croatia, Dominican Republic, Ecuador, Egypt, El Salvador, Hungary, Lebanon, Malaysia, Mexico, Morocco, Nigeria, Pakistan, Panama, Peru, Philippines, Poland, Russia, South Africa, South Korea, Thailand, Tunisia, Turkey, Ukraine, Uruguay, and Venezuela.

10. Brazil, Indonesia, Malaysia, Mexico, South Africa, South Korea, Thailand, and Turkey.

11. Argentina, Australia, Austria, Belgium, Brazil, Bulgaria, China, Colombia, Egypt, France, Germany, Hungary, Indonesia, Ireland, Italy, Japan, Kazakhstan, Korea, Malaysia, Mexico, Netherlands, Peru, Philippines, Poland, Portugal, Russia, South Africa, Spain, Thailand, Turkey, Ukraine, the UK, and the US.

12. Five-year contracts were preferred for reasons of compatibility with the relevant literature (e.g. Akdoğan & Chadwick, 2015; Chan Lau & Kim, 2004).

13. Here and below, L represents the logarithmic transformation of a variable.

14. The first break relates to the 2007–2008 financial crisis, the second to the 2014 Turkish presidential election, the first election in which the president was directly elected by the people, and the third to the 2016 Turkish coup d’état attempt.

15. We repeated asymmetric cointegration and error correction analyses between the bond market and the CDS market on the daily data for three sub-periods, following Ngene et al. (2017). The first period is the pre-crisis period, from 2005 to the 2007 global crisis. The second sub-period is the 2007–2009 crisis period. The third and last sub-period is the post-crisis period, encompassing 2009–2020. The MTAR approach in particular confirms that the cointegration relationship between the series in all three periods is not symmetric but asymmetric, especially in positive discrepancies, in accordance with the first findings. In accord with Ngene et al. (2017), we find that the bidirectional causality relationship is valid between both markets in the short run in the pre-financial crisis and financial crisis periods, but there is a causality relationship from CDS spreads to EMBI+ spreads in the post-crisis period. As for the long-run evaluation, it is determined that the bond market leads the CDS market in price discovery in the pre-crisis period, while the CDS market leads the bond market in the post-crisis and crisis periods. Therefore, while the results obtained from the periodic distinction reveal the leadership role of the CDS market in price discovery especially in the post-crisis period, they also suggest that the liquidity problems that may exist in the bond market may be in question especially in the post-crisis period.