Cross-Sectional ARDL
Cross-Sectional Autoregressive Distributed Lag · Also known as: Panel ARDL with cross-sectional dependence
CS-ARDL (Cross-Sectional ARDL) applies the ARDL framework to panel data while explicitly accounting for cross-sectional dependence—correlation of shocks and relationships across units (countries, firms, regions). Introduced by Pesaran and colleagues (2016), it extends panel ARDL methods to handle common factors or global shocks affecting all units simultaneously. This is crucial for realistic modeling of internationally integrated economies and firm networks.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use CS-ARDL when analyzing internationally linked data (multiple countries), firm networks, or any setting with presumed common shocks. It is essential for macroeconomic panels (countries/regions) and financial panels (banks, firms) where global factors matter. Assume long time series (T > 20) and reasonable panel length (N > 10).
Strengths & limitations
- Explicitly models cross-sectional dependence via common factors
- Yields consistent estimates even when units are cross-sectionally dependent
- Allows heterogeneous long-run and short-run coefficients across units
- Natural inference on cross-sectional variation in responses to shocks
- Requires specification of number of common factors; under-specification biases, over-specification reduces efficiency
- Estimation more complex than standard panel ARDL; requires specialized software
- Assumes weak cross-sectional dependence (factors are few); strong dependence may require more factors
- Inference on common factors themselves is not straightforward
Frequently asked
How do I determine the number of common factors?
Use information criteria (Bai-Ng ICp, ICp2) or eigenvalue ratios of cross-sectional averages. Cross-validate by testing robustness to factor number plus/minus 1.
What is the difference between CS-ARDL and standard panel ARDL?
Standard panel ARDL ignores cross-sectional dependence; CS-ARDL models it via common factors. If dependence is weak or truly absent, both yield similar results. If strong dependence exists, standard ARDL biases estimates and overstates significance.
Can I identify economic shocks from common factors?
Not directly; latent factors lack inherent labels. Examine correlations of extracted factors with observable macro variables (global GDP growth, commodity prices, financial stress indices) to infer their meaning.
How do heterogeneous coefficients across units arise in CS-ARDL?
Units respond differently to the same shock due to structural differences (openness, sector composition, initial conditions). CS-ARDL captures this while controlling for common shocks; coefficient heterogeneity reflects structural diversity, not data artifacts.
Sources
- Pesaran, M. H., & Smith, R. (2016). Testing weak cross-sectional dependence in large panels. Econometric Reviews, 34(6-10), 1089-1117. link ↗
- Chudik, A., Kapetanios, G., & Pesaran, M. H. (2018). A one covariate at a time, multiple testing approach to variable selection in high-dimensional linear regression models. Econometric Reviews, 37(8), 953-1010. DOI: 10.3982/ecta14176 ↗
How to cite this page
ScholarGate. (2026, June 3). Cross-Sectional Autoregressive Distributed Lag. ScholarGate. https://scholargate.app/en/econometrics/cs-ardl
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
Compare side by side →