Glossary

Phillips-Perron Test

A unit-root test that assesses whether a time series is stationary. The PP test uses a non-parametric correction for serial correlation and heteroscedasticity rather than adding lagged differences.

Definition

A unit-root test that assesses whether a time series is stationary. The PP test uses a non-parametric correction for serial correlation and heteroscedasticity rather than adding lagged differences.

Why It Matters

Unlike the Augmented Dickey-Fuller test, which addresses serial correlation by adding lagged difference terms, the Phillips-Perron test corrects for it non-parametrically. This makes the PP test less sensitive to the choice of lag length, which can be a source of ambiguity in ADF testing. The PP test is particularly useful when the researcher suspects that the lag structure of the ADF test may be misspecified, and it serves as a complementary check alongside the ADF test in robust unit-root analysis.

Example

A researcher testing whether the Turkish real exchange rate contains a unit root applies both the ADF test (with 4 lags) and the PP test to the same series. The ADF test fails to reject the null at the 5% level (p = 0.08), but the PP test rejects it (p = 0.03). The discrepancy highlights the sensitivity of ADF results to lag selection and reinforces the value of using multiple unit-root tests.

Related Terms

Software Notes

SPSS: Not natively available; use R integration for the Phillips-Perron test.
R: Use PP.test() in base R or ur.pp() from the urca package for the Phillips-Perron test with detailed output and critical values.
Stata: Use pperron variable to perform the Phillips-Perron test. Add , trend to include a trend term or , noconstant to exclude the constant.