Volatility transmission across currencies and stock markets: GIIPS in crisis

Abstract

This article explores the structure of the volatility transmission mechanism between stock and currency markets for Eurozone economies with systemic fiscal problems such as Greece, Italy, Ireland, Portugal and Spain. We focus on the structural properties of volatility diffusion, in times of market instability, illiquidity, fiscal crisis as well as political uneasiness. Our evidence indicates the presence of bidirectional, asymmetric volatility spillovers between currency and stock markets. Our empirical findings bear significant implications for the allocation of money and capital in diversified investment choices, in a global marketplace of erratic restructurings and systemic instability.

Keywords:

• stock market
• foreign exchange market
• EGARCH
• volatility spillover
• Eurozone

JEL Classification:

• G15
• G10

Notes

¹ With respect to alternative volatility models: if we were to use the GJR GARCH, for example, we would have to impose parameter restrictions in order to ensure a positive sign for volatility. We also did not use the QGARCH specification because of evidence that it underpredicts volatility which is associated with bad news (e.g. Engle and Ng, 1993; bad news are typical of our sample period). One drawback of model-based methods (such as those based on EGARCH), in the analysis of contagion and spillover effects, is that they do not make a distinction between structural association among markets and contagion effects (Baele, 2005). Moreover, these approaches can be problematic in the case of common unobservable shocks across markets (Bekaert et al., 2005). Nevertheless, such concerns are not prohibitive to our modelling approach because (a) a small open economy like Greece of Portugal does not exert systemic influence over the market of, say, the Chinese yuan or the Swiss franc, and (b) shocks in Greece and Portugal are not qualitatively ‘common’ with shocks in the US or the Chinese economy: stock market fluctuations in the South of Europe, were – in the context of our data set – originated in systemic macroeconomic problems which are qualitatively distinct from growth dynamics of the Chinese or the US economy. The employment of a GARCH-type model has the advantage of capturing the transmission of volatility across markets, which is a typical of financial crises. Also, a modelling specification like EGARCH allows us to assess the short run dynamics of asset markets and specifically their sensitivity to the size and the sign of market innovations.

² As a robustness check, we run our econometric model, applying the MSCI country indices instead of national stock market indices. These findings were not substantially different. Our results are available upon request.