Spillover of tail risk : an application to major european and US stock indices

Abstract

International financial crises occurred during the past two decades have turned academic researcher’s attention to financial spillovers. Previous studies have dealt with and eventually detected return and volatility spillovers across countries during crises periods. However, no previous research has dealt with the spillover of tail risk. In this paper our objective is to detect, if there exists, spillover of tail risk using a well known measure for risk management purposes; Value at Risk. We calculate the 1 day, 95% and 99% Value at Risk for major European and US stock indices - CAC40 (France), DAX30 (Germany), FTSE100 (UK), EUROSTOXX50 (Europe) and DJ INDUSTRIALS, NASDAQ100 and S&P500 (USA). We follow the historical simulation and variance approach; variance is estimated as a moving average, an exponentially weighted moving average and a GARCH-type model. We also use Extreme Value Theory as an alternative method to calculate VaR. In order to investigate spillover effects, Granger causality and contemporaneous and lagged relationships across the changes in the VaR of the various indices are examined, since the VaR series in levels are observed as non stationary. We also attempt to capture the concept of cointegration of the non stationary series examined. The results of the current research indicate that spillovers of tail risk do exist; US indices causes European markets, but Europe as well - especially FTSE100 - has effects on USA. In addition, regional effects and also contemporaneous effects are observed. These results help the forecast of the behaviour of a market, when information for a negative shock in another foreign market exists.