GAUK 910836 -- Disentangling permanent and transitory spillovers.
|Principal investigator:||Mgr. Tomáš Křehlík M.A., Ph.D.|
doc. PhDr. Jozef Baruník Ph.D.
Mgr. Luboš Hanus
|Description:||Traditional approach to estimating spillovers is using forecast error variance decomposition from vector autoregression system. In our research, we first critically revisit the current methodology of spillovers, as the spillover estimation is often performed on fractionally co-integrated and long-memory volatility series ignoring these facts. We provide large-scale simulations, which elucidate to what extent is the current spillover research valid under such settings. Second, we propose a general spectral-based framework for disentangling long-term spillovers from short-term ones. Moreover, we offer two economic applications of this methodology. First, we investigate spillover structure of a broad spectrum of financial assets and the effects of crisis. Second, we show implications of proposed methodology for long-term versus short-term portfolio optimization.|
|Work in grant:|
|Finance:||Accepted for financing in April 2014|