Error Correction Models for Fractionally Cointegrated Time Series
This paper provides a proof of Granger's error correction model [see \textit{C. W. J. Granger}, Oxford Bull. Econ. Stat. 48, 213--228 (1986)]) for fractionally cointegrated variables and points out a necessary assumption that has not been noted before. Let \(x_{t}\) be an \(n\)-dimensional vector of an \(I(1)\) process. The process \(x_{t}\) is called fractionally cointegrated (\(x_{t}\sim FCI(d)\)) if there is an \(a\in R^{n}\), \(a\neq0\), such that \(a'x_{t}\sim I(d)\) with \(0<d<1\). If there is a fixed \(\widetilde d\in(0,1)\) such that any linear combination of \(x_{t}\) is either \(I(\widetilde d)\) or \(I(1)\), the process \(x_{t}\) is called purely fractionally cointegrated with cointegration rank \(r\) (\(x_{t}\sim FCI(d,r)\)). The author considers the multivariate World representation of \(\Delta x_{t}\): \((1-B)x_{t}=\delta+\varepsilon_{t}\), where \(C(B)=1_{n}+C_1B+C_2B^2+\ldots\), \(\varepsilon_{t}\) is a white-noise vector process; \(Bx(t)=x(t-1)\). It is proved that if \(x_{t}\) is purely \(FCI(d,r)\) with full-rank \(n\times r\) cointegrating matrix \(A\), then \(A'\delta=0\) and \(A'C(1)=0\). Furthermore, there exist linear filters \(A(B)\) and \(d(B)\) and an \(n\times r\) matrix \(\Gamma\) of rank \(r\) such that \[ A(B)(1-B)x_{t}=-[1-(1-B)^{1-d}](1-B)^{d}\Gamma z_{t}+d(B)\varepsilon_{t} \] with \(z_{t}=A'x_{t}\) and \(A(0)=1_{n}\), where \(d(B)\) is a scalar filter. A simpler error correction model is proposed which can be employed to estimate fractionally cointegrated systems in three steps.
- scientific article; zbMATH DE number 3872515
- Co-Integration and Error Correction: Representation, Estimation, and Testing
- scientific article; zbMATH DE number 1031883
- Polynomial cointegration. Estimation and test
- Representation of Cointegrated Autoregressive Processes with Application to Fractional Processes
- Modelling systems with a mixture of \(I(d)\) and \(I(0)\) variables using the fractionally co-integrated VAR model
- Extended complex error correction models for seasonal cointegration
- Modelling structural breaks, long memory and stock market volatility: an overview
- Bias Correction of Persistence Measures in Fractionally Integrated Models
- scientific article; zbMATH DE number 3872515 (Why is no real title available?)
- A representation theory for polynomial cofractionality in vector autoregressive models
- Representation of Cointegrated Autoregressive Processes with Application to Fractional Processes
- A REPRESENTATION THEORY FOR A CLASS OF VECTOR AUTOREGRESSIVE MODELS FOR FRACTIONAL PROCESSES
- First-order bias correction for fractionally integrated time series
- scientific article; zbMATH DE number 1031883 (Why is no real title available?)
- Support for Governments and Leaders: Fractional Cointegration Analysis of Poll Evidence from the UK, 1960-2004
This page was built for publication: Error Correction Models for Fractionally Cointegrated Time Series
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4828168)
