TY  - JOUR
AU  - Ghysels,Eric
AU  - Santa-Clara,Pedro
AU  - Valkanov,Rossen
TI  - Predicting Volatility: Getting the Most out of Return Data Sampled at Different Frequencies
JF  - National Bureau of Economic Research Working Paper Series
VL  - No. 10914
PY  - 2004
Y2  - November 2004
UR  - http://www.nber.org/papers/w10914
L1  - http://www.nber.org/papers/w10914.pdf
N1  - Author contact info:
Eric Ghysels
Department of Economics
University of North Carolina-Chapel Hill
Gardner Hall, CB 3305
Chapel Hill, NC 27599-3305
Tel: 919/966-5325
Fax: 919/966-4886
E-Mail: eghysels@unc.edu

Pedro Santa-Clara
Faculdade de Economia
Universidade Nova de Lisboa
Rua Marques de Fronteira, 20
1099-038 LISBOA
PORTUGAL
Tel: +351-91-493-4313
E-Mail: psc@fe.unl.pt

Rossen Valkanov
UC, San Diego
E-Mail: rvalkanov@ucsd.edu
AB  - We consider various MIDAS (Mixed Data Sampling) regression models to predict volatility. The models differ in the specification of regressors (squared returns, absolute returns, realized volatility, realized power, and return ranges), in the use of daily or intra-daily (5-minute) data, and in the length of the past history included in the forecasts. The MIDAS framework allows us to compare models across all these dimensions in a very tightly parameterized fashion. Using equity return data, we find that daily realized power (involving 5-minute absolute returns) is the best predictor of future volatility (measured by increments in quadratic variation) and outperforms model based on realized volatility (i.e. past increments in quadratic variation). Surprisingly, the direct use of high-frequency (5-minute) data does not improve volatility predictions. Finally, daily lags of one to two months are sucient to capture the persistence in volatility. These findings hold both in- and out-of-sample.
ER  -