Long memory and stochastic trend
In this paper, we study a general stochastic trend model and provide conditions on the partial sums which imply the convergence of the V/S statistic. These conditions generalize those in Giraitis et al. (J. Appl. Probab. 38 (2001) 1033) obtained in the case of deterministic trend model. As a particular example of stochastic trend we study a regime switching process called mixture model. We prove that in the non-trivial cases the partial sums converge to a compound Poisson process whereas in "degenerated" cases it resembles the behavior of the I(d-1) process.
| Year of publication: |
2003
|
|---|---|
| Authors: | Leipus, Remigijus ; Viano, Marie-Claude |
| Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 61.2003, 2, p. 177-190
|
| Publisher: |
Elsevier |
| Keywords: | Long memory Stochastic trend Compound Binomial process Compound Poisson process Mixture model |
Saved in:
Saved in favorites
Similar items by person
-
Stability of random coefficient ARCH models and aggregation schemes
Kazakevicius, Vytautas, (2004)
-
Stability of random coefficient ARCH models and aggregation schemes
Kazakevičius, Vytautas, (2004)
-
Stability of random coefficient ARCH models and aggregation schemes
Kazakevicius, Vytautas, (2004)
- More ...