出版社：Suntory Toyota International Centre for Economics and Related Disciplines
摘要：Semiparametric estimates of long memory seem useful in the analysis oflong financial time series because they are consistent under muchbroader conditions than parametric estimates. However, recent largesample theory for semiparametric estimates forbids conditionalheteroscedasticity. We show that a leading semiparametric estimate, theGaussian or local Whittle one, can be consistent and have the samelimiting distribution under conditional heteroscedasticity as underconditional homoscedasticity assumed by Robinson (1995a). Indeed,noting that long memory has been observed in the squares of financialtime series, we allow, under regularity conditions, for conditionalheteroscedasticity of the general form introduced by Robinson (1991)which may include long memory behaviour for the squares, such as thefractional noise and autoregressive fractionally integrated movingaverage form, as well as standard short memory ARCH and GARCHspecifications
