Fisher Panel Unit-Root Test
Maddala-Wu (Fisher-type) Panel Unit-Root Test · Also known as: Maddala-Wu Test, Fisher-type Panel Unit-Root Test, MW Panel Unit-Root Test, Fisher Panel Birim Kök Testi
The Fisher-type (Maddala-Wu) panel unit-root test, introduced in 1999, combines individual-level ADF unit-root p-values using Fisher's chi-squared meta-analytic framework to produce a single panel-level test statistic. Unlike the Levin-Lin-Chu approach, it does not impose a common autoregressive parameter across cross-sections, making it a natural choice for heterogeneous panels in macroeconomics, finance, and regional economics.
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 Fisher panel unit-root test when working with unbalanced or balanced panels where the autoregressive parameter is expected to differ across units (heterogeneous panels). It is appropriate for macroeconomic country panels, firm-level financial data, and regional datasets where imposing a single common root is implausible. Key assumptions include cross-sectional independence (or weak dependence); if strong cross-sectional correlation is present, second-generation tests such as CIPS are preferable. Minimum time-series length per unit should be sufficient for reliable ADF inference, typically T >= 20.
Strengths & limitations
- Allows heterogeneous autoregressive parameters across cross-sections, unlike Levin-Lin-Chu.
- Works with unbalanced panels because individual p-values are computed unit by unit.
- Non-parametric combination step is robust to different individual test specifications.
- Asymptotic chi-squared distribution is easy to compute without simulation.
- Maintains cross-sectional independence assumption, which is often violated in practice.
- Rejection identifies at least one stationary unit but does not reveal how many or which ones.
- Power can be low when N is small or only a minority of units are stationary.
- Relies on asymptotic T approximations; small-T panels may yield unreliable p-values.
Frequently asked
How does the Fisher test differ from the Im-Pesaran-Shin (IPS) test?
Both allow heterogeneous autoregressive coefficients, but IPS standardizes individual ADF t-statistics and averages them, relying on tabulated critical values. The Fisher test combines p-values via the chi-squared formula, making it applicable even when individual test distributions are non-standard or when different unit-root tests (ADF or PP) are mixed across units.
Can the Fisher test handle unbalanced panels?
Yes. Because the combination step uses only the p-value from each unit's regression, units with different time-series lengths contribute equally. Each unit's ADF is estimated on its own available observations, and the resulting p-value enters the Fisher sum regardless of whether Ti matches other units.
What should I do if cross-sectional dependence is detected?
If a CD test (e.g., Pesaran 2004) signals significant cross-sectional dependence, the Fisher test's size may be distorted. In that case, second-generation panel unit-root tests such as the CIPS test (Pesaran 2007), which explicitly account for common factors, are more appropriate alternatives.
Sources
- Maddala, G. S., & Wu, S. (1999). A comparative study of unit root tests with panel data and a new simple test. Oxford Bulletin of Economics and Statistics, 61(S1), 631–652. DOI: 10.1111/1468-0084.0610s1631 ↗
How to cite this page
ScholarGate. (2026, June 2). Maddala-Wu (Fisher-type) Panel Unit-Root Test. ScholarGate. https://scholargate.app/en/econometrics/fisher-panel-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 TestEconometrics↔ compare
- Im-Pesaran-Shin TestEconometrics↔ compare
- Levin-Lin-Chu TestEconometrics↔ compare