۱- Introduction

۲- Methodology

۳- Data

۴- Results

۵- Conclusions

References

Abstract

This paper analyses the co-movements and contagion between 24 European Union stock markets. We apply a Dynamic Conditional Correlation – Mixed Data Sampling model to extract short and long-run correlations. We use short-term correlations to detect contagion. Finally, we employ a gravity-type regression to investigate the determinants of long-term correlations. First, we find significant differences between the stock market co-movements, which seem to depend on economic development and market deepening. Moreover, the time varying correlations emphasize different phases of development, i.e. integration, contagion, and divergence. Second, the contagion estimates reveal that, during some crisis episodes, the contagion is temporary, while for other periods the contagion becomes more persistent, indicating a herding behavior. Third, the co-movements determinants show that global factors and economic similarities are important in explaining correlations. Finally, our findings on long-term correlations drivers in contagion times are mixed, revealing, on the one hand, a pure contagion that is not explained by fundamentals, and, on the other hand, a wake-up call in terms of cross-border bank flows.

Introduction

The global financial crisis (GFC) and the European sovereign debt crisis (ESDC) accentuated the changing nature of the financial markets. Moreover, in European Union (EU), the measures implemented by authorities to strengthen the single market and the presumed integration paths have influenced the financial markets. Given that the financial markets became very fickle over the last decade, we consider of great relevance to study the patterns and determinants of the co-movements and contagion in EU stock markets. In this paper, our objective is threefold. First, we document the time varying nature of the pairwise dynamic conditional correlation (DCC) for 24 stock markets. We group the markets into developed, emerging, and frontier and we intend to identify integration paths and crisis effects. Second, we investigate the contagion incidence during the turmoil period. The definition and the measurement of contagion are widely debated in the literature (Forbes and Rigobon, 2002; Bekaert et al., 2014; Caporin et al., 2013; Dungey and Gajurel, 2014). In our approach, we follow Forbes and Rigobon (2002) who define contagion as a significant increase in the level of co-movements between two markets after a crisis. Moreover, if investors from different markets show a simultaneous behavior when the markets co-movements coefficients are high, then this is a sign of herding behavior (Chiang et al., 2007).