Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
9549560 | Economics Letters | 2005 | 7 Pages |
Abstract
Semiparametric estimation of the memory parameter in economic time series raises the problem of the small sample size and the poor approximation of the asymptotic distribution to the finite sample counterpart. This paper considers the bootstrap to improve the finite sample distribution of the popular log peridogram regression and shows that it can significantly reduce the error in the coverage rates of the confidence intervals.
Related Topics
Social Sciences and Humanities
Economics, Econometrics and Finance
Economics and Econometrics
Authors
J. Arteche, J. Orbe,