Phillips-Perron Unit Root Test

The Phillips-Perron test checks the null hypothesis of a unit root by correcting the Dickey-Fuller t statistic non-parametrically for serial correlation with a Newey-West long-run variance.

Phillips-Perron 검정은 Newey-West 장기분산으로 Dickey-Fuller t 통계량의 계열상관을 비모수적으로 보정해 단위근 귀무가설을 검정하는 방법이다.

Is my interest-rate series stationary, or does it need differencing?

금리 시계열이 정상인지, 차분이 필요한지 알고 싶어요.

The test estimates a Dickey-Fuller regression without extra lags and adjusts the t statistic using the long-run variance of the residuals. Rejection of the null indicates the series is stationary around its deterministic terms.

When to use Phillips-Perron Unit Root Test

• You want a unit root test robust to unspecified serial correlation and heteroskedasticity.
• You want a check alongside the ADF test.
• You are deciding whether to difference before modelling.

When to use something else

• The null of stationarity is preferred; use the KPSS Test.
• The series has large negative moving-average components, where PP over-rejects; use the Augmented Dickey-Fuller Test.

Assumptions and requirements

• Correct deterministic terms (none, constant, or constant and trend).
• Adequate bandwidth for the Newey-West estimator.
• No structural breaks in the sample.

What the analysis reports

• Z(τ) statistic and MacKinnon p-value
• Critical values at 1%, 5% and 10%
• Number of Newey-West lags

How to report it

A Phillips-Perron test with a constant gave Z(τ) = [value], p = [value]; the unit-root null was [rejected/not rejected].

Related methods

• Augmented Dickey-Fuller Test
• KPSS Test
• Engle-Granger Cointegration Test

Run Phillips-Perron Unit Root Test on Quorant