A statistical procedure for testing financial contagion

open2noThe analysis of the relationships among financial markets and the identification of financial contagion episodes are relatively recent in the economic analysis and have experienced a rapid development in the last decade, coinciding with the occurrence of relevant financial crises which had effects that spread outside the geographical areas where they originally started. The increasing interest in this topic has lead to the definition of different tests for detecting the existence of financial contagion (Corsetti et al., 2001; Forbes and Rigobon, 2001; Dungey et al., 2004; Allen and Gale, 2005; Rodriguez, 2007; Krishnamurthy, 2009; Sugihara, 2010). However, conclusions on both theoretical and statistical analyses of financial contagion are far from unique. The changes in the international dynamics of returns, which in the last decades has been characterized by increases in both volatilities and asset price synchronicities in different countries, have raised even further the scientific interest in this topic. In this paper, we propose a new methodology for the evaluation of contagion based on the extent of disequilibria in financial dynamics and, in this framework, we define an innovative test for the detection of contagion which specifically identifies the disequilibrium originated by the international transmission of financial crises and their relationships with the behaviours of market participants. Disequilibria exogenously generated by the spread of the effects of a crisis beyond the dynamic process describing endogenous amplification of volatility from one country to other countries are attributed to contagion phenomena. In this framework, contagion effects are separated from the endogenous transmission processes which have their genesis in both the pricing process system and the investor’s behaviours and which are responsible for the amplification of cross-market financial interdependence. In this paper, we discuss the theoretical framework underlying our approach and define a new econometric model for evaluating contagion among countries.openA. Gardini; L. De AngelisA. Gardini; L. De Angeli