Hedge funds, pension funds, mutual funds, and banks primarily use sovereign credit default swaps (SCDS) to manage country or national credit risk. As a proxy for sovereign risk, SCDS are assumed to have a high correlation with their respective bond and equity markets. Prior researchers have focused mostly on how emerging markets as an asset class are affected by stock and bond markets in times of distress, with scant attention paid to either post-crisis periods or the interconnectedness of credit risk across emerging markets. Spillover or contagion is expected during times of crisis because financial asset prices tend to move together at such times. Thus, crisis-focused research is not necessarily useful for providing diversification insights. Identifying regional differences in the commonality of SCDS correlations across a variety of market conditions affords an opportunity for enhanced diversification and risk reduction.

Dynamic conditional correlation (DCC), impulse response function (IRF) analysis, and principal components analysis (PCA) reveal that individual Asian emerging markets are significantly linked with one another during and after a crisis but are relatively unaffected by developments in other regions’ emerging markets. Developed markets are found to have a greater impact on Asian emerging markets. Sovereign risk linkage is greater for European and Latin American emerging markets in times of crisis than in normal times.

The diversification benefits of using SCDS in Asian emerging markets are muted because of high correlations, strong commonality, and higher persistence of shocks. Correlations are lower across European and Latin American emerging markets during normal periods, providing significant diversification opportunities. Interregional markets exhibit the lowest DCC and the least commonality.

The sample includes 15 emerging markets organized into three regions: Asia (China, Indonesia, Malaysia, the Philippines, and Thailand), Europe (Russia, the Czech Republic, Hungary, Poland, and Turkey), and Latin America (Brazil, Argentina, Chile, Colombia, and Mexico). The authors use daily five-year SCDS spread data from Datastream, covering 3 March 2008 to 27 October 2014. The sample is broken down into crisis and post-crisis periods, taking into account the US subprime crisis and the European debt crisis. The authors define the crisis period to be from March 2007 to August 2012, but their data start in March 2008.

They use three approaches in studying relationships across emerging markets. To detect how SCDS return series are correlated over time, the authors use a DCC model and a generalized autoregressive conditional heteroskedasticity (GARCH) model to approximate univariate parameters for each return series; the results are used to estimate a time-varying correlation matrix. To determine how shocks to the volatility of other series affect the volatility of a selected series, they use a multivariate GARCH model (a diagonal BEKK formulation) to estimate the respective volatilities. The authors then use an IRF from an estimated vector autoregression (VAR) model to determine the degree of volatility shock across pairs of country SCDS. PCA is used to determine any common factors that may influence SCDS return-series correlations.

The authors’ findings are useful for institutional investors with the requisite assets and expertise to create specialized SCDS portfolios across regions. Smaller investors who consider the insights from this research when structuring portfolios will be better prepared for how individual and regional emerging markets may react in both normal times and times of stress.