Non‑linear Cointegration Test, Based on Record Counting Statistic
This paper proposes a new procedure for cointegration tests, as traditional tests fail to detect the presence of nonlinearities in cointegrated series. The two-step Engle and Granger (EG) test is modified by incorporating the RUR and FB-RUR tests of Aparicio et al. (2006). These non-parametric tests...
| Autores: | , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2023 |
| País: | España |
| Institución: | Universidad Rey Juan Carlos |
| Repositorio: | BURJC-Digital. Repositorio Institucional de la Universidad Rey Juan Carlos |
| OAI Identifier: | oai:burjcdigital.urjc.es:10115/39778 |
| Acceso en línea: | https://hdl.handle.net/10115/39778 |
| Access Level: | acceso embargado |
| Palabra clave: | Cointegration test Monte Carlo Time Series Error Correction Model |
| Sumario: | This paper proposes a new procedure for cointegration tests, as traditional tests fail to detect the presence of nonlinearities in cointegrated series. The two-step Engle and Granger (EG) test is modified by incorporating the RUR and FB-RUR tests of Aparicio et al. (2006). These non-parametric tests, based on functions of order statistics, exhibit desirable properties such as invariance to nonlinear transformations of the series and robustness to significant parameter shifts. Furthermore, no prior estimation of the cointegrating parameter is required, resulting in parameter-free asymptotic null distributions. Monte Carlo simulations are used to evaluate the test’s properties and power at different sample sizes, demonstrating their ability to detect cointegration in real exchange rate relationships where standard cointegration tests fail. |
|---|