Abstract
We compare three methods of constructing confidence intervals for sample autocorrelations of squared returns modeled by models from the GARCH family. We compare the residual bootstrap, block bootstrap and subsampling methods. The residual bootstrap based on the standard GARCH(1,1) model is seen to perform best. © Springer-Verlag Berlin Heidelberg 2004.
Cite
CITATION STYLE
Kokoszka, P., Teyssière, G., & Zhang, A. (2004). Confidence intervals for the autocorrelations of the squares of GARCH sequences. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3039, 827–834. https://doi.org/10.1007/978-3-540-25944-2_107
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
