Dealing with complex time series data requires careful attention to unit roots and cointegration — issues addressed by this article.
Data & Methods:
This work advocates for the autoregressive distributed lag (ARDL) model as a practical solution. It relies on bounds testing procedures, demonstrated through extensive Monte Carlo simulations showing improved performance across various scenarios.
Key Findings:
The ARDL approach proves more effective than traditional methods in multiple common situations involving time series data integration and unit roots.
Why It Matters:
Researchers can now implement this technique with clearer examples. Step-by-step replication guides are included, along with specialized software designed to test for cointegration and dynamically model results using the ARDL framework.
This piece provides an accessible pathway for political scientists navigating non-stationary time series data complexities.
| Have Your Cake and Eat it Too? Cointegration and Dynamic Inference from Autoregressive Distributed Lag Models was authored by Andrew Philips. It was published by Wiley in AJPS in 2018. |
