A note on testing for nonstationarity in autoregressive processes with level dependent conditional heteroskedasticity

被引：0

作者：

Paulo M. M. Rodrigues

Antonio Rubia

机构：

[1] University of Algarve,Faculty of Economics

[2] University of Alicante,Department of Financial Economics

来源：

Statistical Papers | 2008年 / 49卷

关键词：

Maximum likelihood estimation; Nonstationarity; Volatility; Interest rates;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

In this paper, we investigate the empirical distribution and the statistical properties of maximum likelihood (ML) unit-root t-statistics computed from data sampled from a first-order autoregressive (AR) process with level-dependent conditional heteroskedasticity (LDCH). This issue is of particular importance for applications on interest rate time-series. Unfortunately, the extent of the technical complexity related associated to LDCH patterns does not offer a feasible theoretical analysis, and there is no formal knowledge about the finite-sample size and power behaviour or the ML test for this context. Our analysis provides valuable guidelines for applied work and directions for future work.

引用

页码：581 / 593

页数：12