Fourier Zivot-Andrews Unit Root Test
Fourier-Approximation Zivot-Andrews Unit Root Test · Also known as: Fourier ZA test, FZA unit root test, Fourier structural break unit root test, smooth structural break ADF test
The Fourier Zivot-Andrews test extends the classic Zivot-Andrews (1992) unit root test by replacing sharp, single structural break dummies with a low-frequency Fourier approximation, allowing the test to accommodate smooth, gradual, and multiple unknown breaks in the level or trend of a series.
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 Fourier Zivot-Andrews test when you suspect a time series is subject to smooth or gradual structural change at an unknown date — common in macroeconomic aggregates, commodity prices, financial series, and climate data. It is preferable to standard ADF or Zivot-Andrews when visual inspection or economic reasoning suggests the break was not instantaneous. Do not use it as a routine replacement for ADF in the absence of structural change suspicion, because including unnecessary Fourier terms can reduce power. It is also not designed for multiple sharp breaks; consider Bai-Perron or other multiple-break tests in that case.
Strengths & limitations
- Accommodates smooth, gradual, and multiple structural breaks without requiring the researcher to pre-specify break dates or functional forms.
- More powerful than standard ADF and conventional Zivot-Andrews when breaks are genuinely smooth rather than sharp.
- Low-frequency Fourier terms provide a parsimonious approximation that avoids overfitting.
- Robust to misspecification of the break type — sine and cosine terms can mimic many different break shapes.
- Builds on the well-established Zivot-Andrews framework, making it conceptually accessible to practitioners familiar with structural-break unit root testing.
- Critical values are non-standard and must be sourced from simulation tables; they depend on the Fourier frequency selected, the model specification, and the sample size.
- Power can be low against alternatives with very sharp or abrupt breaks, where the Zivot-Andrews or Lumsdaine-Papell test may perform better.
- The optimal frequency selection adds a pre-testing step that introduces additional uncertainty and can affect size in finite samples.
- Not designed to estimate or locate the structural break precisely — it is a unit root test, not a break-date estimator.
Frequently asked
How does this differ from the standard Zivot-Andrews test?
The classic Zivot-Andrews test allows for one sharp, instantaneous break in the level or trend, selected by minimising the t-statistic over all possible break dates. The Fourier variant replaces the break dummy with low-frequency sine and cosine terms, so it can handle smooth, gradual, or multiple breaks without locating a single break point.
Which Fourier frequency k should I choose?
Estimate the regression for k = 1, 2, and 3 and choose the k that minimises the sum of squared residuals. Economic series rarely require k > 3; higher frequencies risk overfitting the deterministic component.
Where do I get the critical values?
Critical values are tabulated through simulation in Enders and Lee (2012) for the three standard model specifications (break in level only, break in trend only, and break in both). Use the table matching your sample size and selected k.
Can the Fourier Zivot-Andrews test handle more than one structural break?
Yes, implicitly. A Fourier series with a single low frequency can approximate multiple smooth transitions simultaneously, which is one of its advantages over the single-break Zivot-Andrews test. However, if breaks are sharp and numerous, a dedicated multiple-break test (e.g., Bai-Perron) is more appropriate.
What should I do after the test rejects the unit root null?
Rejection means the series is consistent with trend stationarity with smooth structural change. Include the Fourier terms as deterministic regressors in subsequent models (e.g., ARDL, VAR) rather than differencing the series, to avoid unnecessary information loss.
Sources
- Enders, W., & Lee, J. (2012). A unit root test using a Fourier series to approximate smooth breaks. Oxford Bulletin of Economics and Statistics, 74(4), 574-599. DOI: 10.1111/j.1468-0084.2011.00662.x ↗
- Zivot, E., & Andrews, D. W. K. (1992). Further evidence on the great crash, the oil-price shock, and the unit-root hypothesis. Journal of Business and Economic Statistics, 10(3), 251-270. DOI: 10.1080/07350015.1992.10509904 ↗
How to cite this page
ScholarGate. (2026, June 3). Fourier-Approximation Zivot-Andrews Unit Root Test. ScholarGate. https://scholargate.app/en/econometrics/fourier-zivot-andrews-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
- Fourier ADF unit root testEconometrics↔ compare
- Fourier KPSS testEconometrics↔ compare
- Phillips-Perron unit root testEconometrics↔ compare
- Structural Break ADF Unit Root TestEconometrics↔ compare
- Zivot-Andrews Structural Break TestEconometrics↔ compare