Time Series Regression with a Unit Root

From MaRDI portal





Autoregressive time series with a unit root have attracted much attention in the recent literature from both sides, users and theorists as well. The test statistics used to test for unit roots are based on the rather strong assumption that the sequence of innovation is independent with common variance or that the innovation are i.i.d. \((0,\sigma^ 2)\) or even that they are i.i.d. \(N(0,\sigma^ 2).\) To avoid such strong assumptions the author develops a test for unit roots which allows for quite general weakly dependent and heterogeneously distributed innovations. This test is derived by use of functional limit theory on metric spaces. An interesting feature of these test statistics is that their limiting distributions are identical to those suggested in the literature under the assumption that the innovations are i.i.d. A second issue of the paper is to present a new limiting distribution theory which is based on the concept of continuous data recording. This theory together with an asymptotic expansion given in the paper helps to explain the experimental results obtained by \textit{G. B. A. Evans} and \textit{N. E. Savin} [ibid. 49, 753-779 (1981; Zbl 0468.62021) and ibid. 52, 1241-1269 (1984; Zbl 0579.62014)] on unit roots.




Cited in
(only showing first 100 items - show all)








This page was built for publication: Time Series Regression with a Unit Root

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4720609)