RT Generic
T1 Measuring causality between volatility and returns with high-frequency data
A1 Dufour, Jean-Marie
A1 García, René
A1 Taamouti, Abderrahim
A2 Universidad Carlos III de Madrid. Departamento de Economía, 
AB We use high-frequency data to study the dynamic relationship between volatility and equityreturns. We provide evidence on two alternative mechanisms of interaction between returns andvolatilities: the leverage effect and the volatility feedback effect. The leverage hypothesis assertsthat return shocks lead to changes in conditional volatility, while the volatility feedback effecttheory assumes that return shocks can be caused by changes in conditional volatility through atime-varying risk premium. On observing that a central difference between these alternativeexplanations lies in the direction of causality, we consider vector autoregressive models ofreturns and realized volatility and we measure these effects along with the time lags involvedthrough short-run and long-run causality measures proposed in Dufour and Taamouti (2008), asopposed to simple correlations. We analyze 5-minute observations on S&P 500 Index futurescontracts, the associated realized volatilities (before and after filtering jumps through thebispectrum) and implied volatilities. Using only returns and realized volatility, we find a weakdynamic leverage effect for the first four hours at the hourly frequency and a strong dynamicleverage effect for the first three days at the daily frequency. The volatility feedback effectappears to be negligible at all horizons. By contrast, when implied volatility is considered, avolatility feedback becomes apparent, whereas the leverage effect is almost the same. Weinterpret these results as evidence that implied volatility contains important information onfuture volatility, through its nonlinear relation with option prices which are themselves forwardlooking.In addition, we study the dynamic impact of news on returns and volatility, againthrough causality measures. First, to detect possible dynamic asymmetry, we separate goodfrom bad return news and find a much stronger impact of bad return news (as opposed to goodreturn news) on volatility. Second, we introduce a concept of news based on the differencebetween implied and realized volatilities (the variance risk premium) and we find that a positivevariance risk premium (an anticipated increase in variance) has more impact on returns than anegative variance risk premium.
SN 2340-5031
YR 2008
FD 2008-09
LK https://hdl.handle.net/10016/2991
UL https://hdl.handle.net/10016/2991
LA eng
DS e-Archivo
RD 31 may. 2024