Cambridge University Press eBooks· 2012· Book chapter
An Autoregressive Distributed-Lag Modelling Approach to Cointegration Analysis
- 3,942citations
- 2012year
Short summary
This paper demonstrates that appropriately augmented Autoregressive Distributed Lag (ARDL) models can provide super-consistent estimators for long-run cointegrating relationships, even with I(1) variables, allowing valid inferences using standard asymptotic theory.
AI-generated from the title and abstract; the full text is not read.
Key points
- Augmented ARDL models yield super-consistent estimators for long-run cointegrating coefficients.
- Valid statistical inferences on long-run parameters can be made using standard normal asymptotic theory.
- ARDL consistently estimates long-run coefficients regardless of whether regressors are I(1) or I(0).
- Monte Carlo experiments favor the ARDL approach over fully modified OLS for cointegration analysis.
AI-generated from the title and abstract; the full text is not read.
Abstract
This paper examines the use of autoregressive distributed lag (ARDL) models for the analysis of long-run relations when the underlying variables are I(1). It shows that after appropriate augmentation of the order of the ARDL model, the OLS estimators of the short-run parameters are p T-consistent with the asymptotically singular covariance matrix, and the ARDL-based estimators of the long-run coe¢cients are super-consistent, and valid inferences on the long-run parameters can be made using standard normal asymptotic theory. The paper also examines the relationship between the ARDL procedure and the fully modi…ed OLS approach of Phillips and Hansen to estimation of cointegrating relations, and compares the small sample performance of these two approaches via Monte Carlo experiments. These results provide strong evidence in favour of a rehabilitation of the traditional ARDL approach to time series econometric modelling. The ARDL approach has the additional advantage of yielding consistent estimates of the long-run coe¢cients that are asymptotically normal irrespective of whether the underlying regressors are I(1) or I(0).
The authors' abstract, as published at the source. Cambridge University Press eBooks, 2012 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: General Economics, Econometrics and Finance
General Economics, Econometrics and FinanceEconomics, Econometrics and Finance