Panel ADF Unit Root Test
Panel Augmented Dickey-Fuller Unit Root Test · Also known as: Panel ADF test, IPS test, Im-Pesaran-Shin test, panel unit root test
The Panel Augmented Dickey-Fuller (Panel ADF) unit root test extends the classical ADF framework to panel datasets. By pooling information across cross-sectional units it achieves substantially higher power than single-series ADF tests, allowing researchers to determine whether time-series variables are stationary or integrated of order one before modelling long-run relationships.
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 the Panel ADF test when you have panel data with a moderate to large number of cross-sectional units (N ≥ 10) and a time dimension of at least 20–30 periods, and you need to establish the order of integration before fitting cointegration models (VECM, ARDL, Johansen) or regression in levels. It is especially valuable when the individual time dimension is too short for reliable single-series ADF tests. Do not use it as a substitute for cross-sectionally independent data: if strong cross-sectional dependence is present (common factors, global shocks), apply a second-generation panel unit root test such as CIPS (Pesaran 2007) instead, because the standard Panel ADF assumes cross-sectional independence and will over-reject the null when that assumption is violated.
Strengths & limitations
- Substantially higher power than single-series ADF tests when T is short, by pooling N cross-sectional units.
- The IPS variant allows heterogeneous autoregressive coefficients and lag lengths across units, making it flexible for diverse panels.
- Produces a standard normal test statistic with well-known critical values and p-values.
- Applicable to both balanced and, with some care, unbalanced panels.
- Widely implemented in econometric software (Stata, EViews, R) with straightforward interpretation.
- Assumes cross-sectional independence; common shocks or spatial dependence cause severe size distortions.
- The LLC variant imposes a common unit-root coefficient, which is restrictive when panels are heterogeneous.
- Power remains limited when T is very short (T < 15) even with large N.
- Cannot identify which specific cross-sectional units are stationary; the alternative is that at least one unit is stationary, not all.
Frequently asked
What is the difference between the LLC and IPS panel ADF tests?
Levin-Lin-Chu (LLC) pools all cross-sectional units and estimates a single common autoregressive coefficient, imposing homogeneity. Im-Pesaran-Shin (IPS) estimates separate ADF regressions per unit and averages the t-statistics, allowing heterogeneous persistence. IPS is generally preferred for macro panels where countries or firms may differ substantially in how quickly shocks dissipate.
What should I do if the cross-sectional dependence (CD) test is significant?
Standard Panel ADF tests are unreliable under cross-sectional dependence. Switch to second-generation panel unit root tests such as Pesaran's CIPS (Cross-Sectionally Augmented IPS, 2007), which augments each ADF regression with cross-sectional averages to absorb common factors.
If the Panel ADF rejects the null, does that mean all series are stationary?
No. The alternative hypothesis is that at least one (or a non-zero fraction of) cross-sectional unit is stationary. Some units may still contain a unit root. You cannot identify which units are stationary solely from the panel test; inspect individual ADF results for guidance.
How many time periods and cross-sections do I need?
A practical rule of thumb is N ≥ 10 cross-sectional units and T ≥ 20 time periods for the asymptotic approximation to be reliable. With smaller T the test has limited power; with very small N the normal approximation deteriorates.
Should I include a trend in the panel ADF regression?
Only if economic reasoning supports a deterministic trend in the series (e.g. trending GDP per capita). An unnecessary trend inflates critical values and reduces power; a missing trend causes size distortion. Test with and without trend when unsure, and let theory guide the specification.
Sources
- Im, K. S., Pesaran, M. H., & Shin, Y. (2003). Testing for unit roots in heterogeneous panels. Journal of Econometrics, 115(1), 53–74. DOI: 10.1016/S0304-4076(03)00092-7 ↗
- Levin, A., Lin, C.-F., & Chu, C.-S. J. (2002). Unit root tests in panel data: Asymptotic and finite-sample properties. Journal of Econometrics, 108(1), 1–24. DOI: 10.1016/S0304-4076(01)00098-7 ↗
How to cite this page
ScholarGate. (2026, June 3). Panel Augmented Dickey-Fuller Unit Root Test. ScholarGate. https://scholargate.app/en/econometrics/panel-adf-unit-root-test
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.
- Augmented Dickey-Fuller unit root testEconometrics↔ compare
- Panel ARDL Bounds TestEconometrics↔ compare
- Panel Engle-Granger CointegrationEconometrics↔ compare
- Panel Johansen CointegrationEconometrics↔ compare
- Panel KPSS testEconometrics↔ compare
- Panel PP unit root testEconometrics↔ compare