eThesis - Kauppakorkeakoulun sähköiset gradut

OBJECTIVES OF THE STUDY:

The purpose of this study is to examine the drivers behind the time-varying correlation and the volatility linkages between stocks and bonds. Understanding the contemporaneous cross-asset movements is crucial for numerous investment and risk management applications as well as for regulatory policy decisions. I pay special attention to the periods of strong negative stock-bond correlation, since positive long-run relation is widely assumed. In the empirical part, I study if stock or bond market uncertainty, stock market trading activity or inflation expectations explain the somewhat puzzling fluctuating return correlation. Furthermore, I examine which one of the assets is more influential in terms of volatility transmission.

DATA AND METHODOLOGY:

I study daily returns of the Euro Stoxx 50 Index and the German Bund interest rate futures from January 1999 through March 2012 as these assets are one of the most liquid and followed securities globally and hence work as a decent proxy for the stock-bond relationship. I generate two correlation series; the forward-looking 22-trading-day and the Dynamic Conditional Correlation (DCC) based on GARCH-methodology. Multivariate OLS regression is used in the correlation analysis and the multivariate asymmetric full BEKK-GARCH model to examine the volatility linkages.

This thesis shows that stock-bond correlation is considerably time-varying and stock market uncertainty is associated with substantially strong negative return correlation.The probability for negative forward-looking stock-bond correlation is 99 per cent and the mean correlation is -0.56 (the whole sample mean -0.32) when VSTOXX closes higher than 40. I also find stronger negative stock-bond correlation following significant daily VSTOXX changes.

FINDINGS OF THE STUDY:

I find strong evidence for two-way cross-asset volatility spillovers. The whole sample analysis suggests that past stock market volatility affect bonds more heavily than vice versa, but the coefficients are substantially close to each other, or even prefer the Bund futures in the subsample breakdown. It seems that the volatility linkages remain equally strong despite the stock market trend, but the asymmetry of the volatility transmission seems to increase jointly with the stock market uncertainty.