Page 46 - Textos de Matemática Vol. 47
P. 46
36 E. GONC¸ALVES AND C. M. MARTINS
that the functions ⇢|X| and ⇢X2 are almost equal. Otherwise, in the leptokurtic processes, the Taylor property is present in a significant way, becoming stronger when the kurtosis of the model is also stronger.
Again we conclude that the Taylor property is highly dependent on the greater or lesser weight of the tails of the law of the process under study.
Acknowledgments
The authors are grateful to the referee for his valuable comments.
References
[1] E. Gonc¸alves, J. Leite, and N. Mendes-Lopes, A mathematical approach to detect the Taylor property in TARCH processes, Statist. Probab. Lett. 79, 602–610, 2009.
[2] E. Gonc¸alves, C. M. Martins, and N. Mendes-Lopes, Propriedade de Taylor em processos autorregressivos, Estat´ıstica: A ciˆencia da incerteza, Atas do XXI Congresso Anual da Sociedade Portuguesa de Estat´ıstica, I. Pereira, A. Freitas, M. Scotto, M. E. Silva, C. D. Paulino (eds.), Edi¸c˜oes SPE, 51–53, 2014.
[3] E. Gon¸calves, C. M. Martins, and N. Mendes-Lopes, The Taylor property in bilinear models, REVSTAT (in press), 2015.
[4] E. Gon¸calves, N. Mendes-Lopes, I. Dorotoviˇc, J. M. Fernandes, and A. Garcia, North and South Hemispheric Solar Activity for Cycles 21-23: Asymmetry and Conditional Volatility of Plage Regions Areas, Sol. Phys. 289 (6), 2283–2296, 2014.
[5] J. Leite, Processos Threshold GARCH com potˆencia: estrutura probabilista e aplicac¸o˜es a cartas de controlo, Tese de Doutoramento, Universidade de Coimbra, 2013.
[6] M. Haas, Persistence in volatility, conditional kurtosis, and the Taylor property in ab- solute value GARCH processes, Statist. Probab. Lett. 79 (5), 1674–1683, 2009.
[7] C. He and H. Tera¨svirta, Properties of moments of a family of GARCH processes, J. Econometrics 92, 173–192, 1999.
[8] S. J. Taylor, Modelling Financial Time Series, Jonh Wiley & Sons, Chichester, UK, 1986.
[9] H. Tsay and K.S. Chan, A Note on Non-Negative Arma Processes, J. Time Series Anal. 28, 3, 350–360, 2007.
[10] Z. Zhang and H. Tong, On some distributional properties of a first order non-negative bilinear time series model, J. Appl. Probab. 38 (3), 659–671, 2001.
(E. Gon¸calves) CMUC, Department of Mathematics, University of Coimbra, Por- tugal
E-mail address: esmerald@mat.uc.pt
(C. M. Martins) Department of Mathematics, University of Coimbra, Portugal E-mail address: cmtm@mat.uc.pt