recent developments in high frequency financial ... - Index of
recent developments in high frequency financial ... - Index of
recent developments in high frequency financial ... - Index of
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Dynamic modell<strong>in</strong>g <strong>of</strong> large-dimensional covariance matrices 311<br />
correlations exhibit <strong>high</strong> persistence. S<strong>in</strong>ce by <strong>in</strong>corporat<strong>in</strong>g <strong>in</strong>tra-daily <strong>in</strong>formation<br />
these realized measures are also quite precise, this serial dependence can be<br />
exploited for volatility forecast<strong>in</strong>g. A possible extension <strong>of</strong> the methodological<br />
framework suggested <strong>in</strong> the paper could be modell<strong>in</strong>g the realized series <strong>in</strong> a vector<br />
ARMA system, <strong>in</strong> order to analyze volatility spillovers across stocks, <strong>in</strong>dustries<br />
or markets, which however would aga<strong>in</strong> <strong>in</strong>volve a large number <strong>of</strong> parameters.<br />
A closely related area <strong>of</strong> research is concerned with the methods for evaluation<br />
<strong>of</strong> covariance matrix forecasts. In this paper we have used purely statistical evaluation<br />
tools based on a symmetric loss function. An asymmetric measure <strong>in</strong> this case<br />
may have more economic mean<strong>in</strong>g, s<strong>in</strong>ce it is quite plausible to assume that if a<br />
portfolio variance has been overestimated, the consequences are less adverse than<br />
if it has been underestimated. In a multivariate context Byström (2002) uses as an<br />
evaluation measure <strong>of</strong> forecast<strong>in</strong>g performance the pr<strong>of</strong>its generated by a simulated<br />
trad<strong>in</strong>g <strong>of</strong> portfolio <strong>of</strong> ra<strong>in</strong>bow options. The prices <strong>of</strong> such options depend<br />
on the correlation between the underly<strong>in</strong>g assets. Thus the agents who forecast the<br />
correlations more precisely should have <strong>high</strong>er pr<strong>of</strong>its on average.<br />
Further, the models presented <strong>in</strong> this paper can be extended by <strong>in</strong>troduc<strong>in</strong>g the<br />
possibility <strong>of</strong> asymmetric reaction <strong>of</strong> (co)volatilities to previous shocks (leverage).<br />
This can be achieved by <strong>in</strong>troduc<strong>in</strong>g some k<strong>in</strong>d <strong>of</strong> asymmetry <strong>in</strong> Eq. (23), e.g.<br />
by <strong>in</strong>clud<strong>in</strong>g products <strong>of</strong> absolute shocks or products <strong>of</strong> <strong>in</strong>dicator functions for<br />
positivity <strong>of</strong> the shocks.<br />
Acknowledgements F<strong>in</strong>ancial support from the German Science Foundation, Research Group<br />
‘Preis-, Liquiditäts- und Kreditrisiken: Messung und Verteilung’ is gratefully acknowledged. I<br />
am thankful to the referee whose comments significantly improved the overall quality <strong>of</strong> the<br />
paper. I would like to thank W<strong>in</strong>fried Pohlmeier, Michael Lechner, Jens Jackwerth and Günter<br />
Franke for helpful comments. All rema<strong>in</strong><strong>in</strong>g errors are m<strong>in</strong>e.<br />
References<br />
Aït-Sahalia Y, Mykland PA, Zhang L (2005) How <strong>of</strong>ten to sample a cont<strong>in</strong>uous-time process <strong>in</strong><br />
the presence <strong>of</strong> market microstructure noise. Rev F<strong>in</strong>anc Stud 18(2):351–416<br />
Andersen T, Bollerslev T, Christ<strong>of</strong>fersen PF, Diebold FX (2006) Volatility forecast<strong>in</strong>g. In:<br />
Elliott G, Granger C, Timmermann A (eds) Handbook <strong>of</strong> economic forecast<strong>in</strong>g, 1st edition,<br />
chap 15, Elsevier<br />
Andersen TG, Bollerslev T (1998) Answer<strong>in</strong>g the skeptics: yes, standard volatility models do<br />
provide accurate forecasts. Int Econ Rev 39:885–905<br />
Andersen TG, Bollerslev T, Diebold FX, Ebens H (2001a) The distribution <strong>of</strong> stock return<br />
volatility. J F<strong>in</strong>anc Econ 61:43–76<br />
Andersen TG, Bollerslev T, Diebold FX, Labys P (2001b) The distribution <strong>of</strong> exchange rate<br />
volatility. J Am Stat Assoc 96:42–55<br />
Andersen TG, Bollerslev T, Diebold FX, Labys P (2003) Model<strong>in</strong>g and forecast<strong>in</strong>g realized<br />
volatility. Econometrica 71:579–625<br />
Bandi FM, Russell JR (2005) Microstructure noise, realized volatility, and optimal sampl<strong>in</strong>g.<br />
Work<strong>in</strong>g paper, Graduate School <strong>of</strong> Bus<strong>in</strong>ess, The University <strong>of</strong> Chicago<br />
Barndorff-Nielsen OE, Shephard N (2004) Econometric analysis <strong>of</strong> realised covariation: <strong>high</strong><br />
<strong>frequency</strong> based covariance, regression and correlation <strong>in</strong> f<strong>in</strong>ancial economics. Econometrica<br />
72:885–925<br />
Bauwens L, Laurent S, Rombouts J (2006) Multivariate GARCH models: a survey. J Appl Econ<br />
21:79–109<br />
Black F, Litterman R (1992) Global portfolio optimization. F<strong>in</strong>anc Anal J 48(5):28–43<br />
Byström H (2002) Us<strong>in</strong>g simulated currency ra<strong>in</strong>bow options to evaluate covariance matrix<br />
forecasts. J Int F<strong>in</strong>anc Mark Inst Money 12:216–230<br />
Diebold FX, Mariano RS (1995) Compar<strong>in</strong>g predictive accuracy. J Bus Econ Stat 13(3):253–263