Testing for Conditional Heteroscedasticity: Some Monte Carlo Results
For the purpose of testing the adequacy of an ARCH/GARCH model after one has been fitted to the data, many researchers use the Box-Pierce statistic as applied to the squared standardized residuals. Recently, Li and Mark (1994) argued that this procedure may be misleading as the asymptotic distribution of the statistic does not converge to a Ï‡2 distribution. They derived the asymptotic distribution of the correlation coefficients of the squared standardized residuals and proposed some diagnostic tests for the ARCH/GARCH models. In this paper we report some Monte Carlo results for the finite sample performance of their tests and some other commonly used diagnostics.
Journal of Statistical Computation and Simulation
Taylor and Francis
TSE, Yiu Kuen and Zuo, X. L..
Testing for Conditional Heteroscedasticity: Some Monte Carlo Results. (1997). Journal of Statistical Computation and Simulation. 58, 237-253. Research Collection School Of Economics.
Available at: http://ink.library.smu.edu.sg/soe_research/136
This document is currently not available here.