Realized jumps on korean financial markets and explaining credit spreads

Abstract

This paper applies the jump detection method proposed by Tauchen and Zhou (2008), which extends Barndorff-Nielsen and Shephard (2004b, 2006)’s bipower variation estimation, to the Korean equity index market and currency exchange market. Estimation of realized jumps from high-frequency intraday prices reveals that contribution of jump components to total realized variation is higher in the Korean stock market than in the US market. Also, our regression analysis shows that the jump volatility of KORSPI200 futures alone explains 45% of the average credit spread variations of Korean firms. The combination of jump volatility and implied volatility can explain 71% of the credit spread variations, and adding other risk factors such as historical volatility, term-spread and the short rate seems to be redundant. Credit spread changes were explained by jump volatility better than any other explanatory variables although adjusted R-square of the regression was low as it was in previous studies.