Robust Phillips-Perron (PP) Unit Root Test
Robust Phillips-Perron Unit Root Test · Also known as: robust Phillips-Perron test, heteroskedasticity-robust PP test, nonparametric robust unit root test, robust PP
The Robust Phillips-Perron unit root test extends the classical PP test by applying corrections — such as heteroskedasticity-consistent covariance estimation or wild-bootstrap critical values — that maintain valid inference when the error variance of a time series is non-constant or exhibits unconditional heteroskedasticity, conditions under which the standard PP test is severely size-distorted.
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 Robust PP unit root test when you suspect that the time series has episodes of changing volatility — such as financial return series, macroeconomic aggregates spanning crises, or energy prices — and you still need to determine whether the series is stationary. It is particularly appropriate as a robustness check alongside the standard PP or ADF tests when ARCH effects or structural volatility shifts are present. Do not use it as a default replacement for ADF when there is no evidence of heteroskedasticity, since the standard tests are more efficient in homoskedastic settings. It is also not a substitute for structural-break unit root tests (e.g., Zivot-Andrews) when the mean or trend undergoes a discrete shift.
Strengths & limitations
- Maintains correct test size under unconditional heteroskedasticity where the standard PP test over-rejects severely.
- Retains the nonparametric serial-correlation correction of the original PP test, so no lag-selection is needed.
- Wild-bootstrap implementation is straightforward in standard software (R urca/CAViaR, Stata sbcunit) and does not require specifying an error distribution.
- Provides a useful robustness check to validate conclusions from standard PP or ADF tests.
- No additional structural assumptions are imposed; the method is nonparametric in its treatment of both autocorrelation and heteroskedasticity.
- Computationally more demanding than the standard PP test because bootstrap resampling requires many replications (typically 999–4999).
- Power against near-unit-root alternatives can be lower than optimised tests designed specifically for heteroskedastic errors.
- Does not account for deterministic structural breaks in the mean or trend; a Zivot-Andrews or similar test should be used if breaks are suspected.
- The wild-bootstrap variant requires choosing a bootstrap scheme (recursive versus fixed design), and results may differ moderately between schemes.
Frequently asked
How does the Robust PP test differ from the standard PP test?
The standard PP test adjusts only for serial correlation in the errors using a kernel long-run variance estimator. The robust version additionally guards against unconditional heteroskedasticity — for example via wild-bootstrap critical values or an HC-consistent long-run variance — which prevents the severe over-rejection the standard test exhibits when error variance is non-constant.
Should I always prefer the Robust PP over the standard PP test?
Not necessarily. If there is no evidence of heteroskedasticity (e.g., ARCH-LM test is insignificant), the standard PP test or ADF is more powerful. Use the robust variant as a confirmatory or primary test when time-varying variance is detected or strongly suspected.
Is the Robust PP test the same as the nonlinear or structural-break PP test?
No. The robust variant addresses non-constant error variance while keeping the linear, no-break assumption about the deterministic component. Structural-break tests (e.g., Zivot-Andrews) allow the intercept or trend to shift at an unknown date. These are distinct problems requiring different remedies.
How many bootstrap replications are needed?
At least 999 replications are recommended for 5% inference; 4999 is more reliable when the p-value is near the significance threshold. Results should be stable across a range of replication counts before being reported.
Which software supports the Robust PP unit root test?
In R, the pvecm and CAViaR packages, as well as bootstrap wrappers around the tseries PP implementation, can produce robust critical values. In Stata, the sbcunit and xtunitroot commands offer heteroskedasticity-robust options. StatWise provides a guided interface that selects the appropriate variant automatically.
Sources
- Phillips, P. C. B., & Perron, P. (1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335–346. DOI: 10.1093/biomet/75.2.335 ↗
- Cavaliere, G., & Taylor, A. M. R. (2008). Bootstrap unit root tests for time series with nonstationary volatility. Econometric Theory, 24(1), 43–71. DOI: 10.1017/S0266466608080043 ↗
How to cite this page
ScholarGate. (2026, June 3). Robust Phillips-Perron Unit Root Test. ScholarGate. https://scholargate.app/en/econometrics/robust-pp-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
- KPSS TestEconometrics↔ compare
- Nonlinear PP unit root testEconometrics↔ compare
- Phillips-Perron unit root testEconometrics↔ compare
- Robust ADF Unit Root TestEconometrics↔ compare
- Zivot-Andrews Structural Break TestEconometrics↔ compare