Skip to contentScholarGate
LibraryBookshelfDeskReview StudioAssistant
Sign in
On this page
IntuitionHow it worksWhen to use itStrengths & limitationsCommon pitfallsApplicationsFrequently asked🔒 Read the full methodSourcesRelated methods
Cite this pageSpotted an issue on this page? Report or suggest a fix →
Home›Econometrics›Fourier Nonlinear ARDL (Fourier NARDL)
Regression modelEconometrics / time series

Fourier Nonlinear ARDL (Fourier NARDL)

Fourier Nonlinear Autoregressive Distributed Lag Model · Also known as: Fourier NARDL, Fourier nonlinear ARDL, F-NARDL, Fourier asymmetric ARDL

Fourier NARDL extends the Nonlinear ARDL (NARDL) bounds-testing framework by adding Fourier trigonometric terms to the error-correction equation, allowing the model to capture smooth, gradual structural breaks in the long-run relationship without requiring the researcher to know or specify the break date in advance.

ScholarGate
  1. Regression model
  2. v1
  3. 2 Sources
  4. PUBLISHED
Cite this page →
Tools & resources
Download slides
Learn & explore

Read the full method

Members only

Sign in with a free account to read this section.

Sign in

Method map

The neighbourhood of related methods — select a node to explore.

Fourier NARDL
Arellano-Bond GMM estima…Fourier ARDL Bounds TestFourier Engle-Granger co…Fourier Granger CausalityNonlinear ARDLVector Error Correction…

When to use it

Use Fourier NARDL when you have time-series data and theory or preliminary evidence suggests: (1) the long-run relationship is asymmetric — positive and negative changes in the driver affect the outcome differently; and (2) the relationship may have shifted gradually over the sample period due to policy changes, regime transitions, or slow structural evolution. It is appropriate when you cannot pinpoint a specific break date. Do not use it when you have a small sample (fewer than about 60 observations), as adding Fourier terms reduces degrees of freedom noticeably. If asymmetry is not theoretically justified, prefer the standard Fourier ARDL bounds test. If breaks are sharp and well-dated, a Zivot-Andrews or Bai-Perron approach is more parsimonious.

Strengths & limitations

Strengths
  • Simultaneously tests for asymmetric cointegration and accommodates smooth structural change without requiring a known break date.
  • Fourier approximation is flexible: a single low-frequency term can capture a wide variety of smooth, nonlinear deterministic trends.
  • Inherits the NARDL advantage of working with a mix of I(0) and I(1) variables, avoiding the need for pre-testing all variables to the same integration order.
  • Dynamic multipliers provide intuitive visual evidence of the asymmetric adjustment process and its speed.
  • Bounds-testing approach remains valid regardless of whether regressors are strictly I(1) or I(0), making it robust to integration uncertainty.
Limitations
  • Requires a reasonably long time series; adding Fourier components reduces degrees of freedom, making the model unreliable for samples below roughly 60 observations.
  • Selection of the optimal Fourier frequency k via grid search is informal and may over-fit in small samples.
  • The joint null tested by the F-bounds statistic conflates no cointegration with no asymmetry, making it difficult to isolate which restriction drives rejection.
  • Critical values for the Fourier-augmented bounds test differ from standard ARDL tables and must be obtained by simulation; tabulated values for all specifications may not be readily available.
  • Interpretation requires care: apparent asymmetry may partly reflect the Fourier correction rather than genuine behavioural asymmetry.

Frequently asked

How do I choose the Fourier frequency k?

Estimate the model for each integer value of k from 1 to some maximum (typically k ≤ 3 or 5) and choose the value that minimises the residual sum of squares. Some software reports the optimal k automatically.

Can Fourier NARDL be used with panel data?

Yes. Panel extensions of Fourier NARDL exist and have been applied in the literature, though they are less standardised than the single-equation version and require simulation-based critical values appropriate for the panel dimension.

What is the difference between Fourier NARDL and Fourier ARDL?

Fourier ARDL assumes symmetric adjustment — positive and negative changes in the regressor have equal effects. Fourier NARDL relaxes this by decomposing the regressor into positive and negative partial sums, testing for both asymmetry and smooth structural change simultaneously.

Do I still need to test for unit roots before running Fourier NARDL?

The NARDL bounds test is valid for regressors that are I(0) or I(1), so full pre-testing is not required. However, you must confirm that no variable is I(2); standard ADF or PP tests (or their Fourier variants) suffice for this check.

How do I interpret the dynamic multipliers?

Dynamic multipliers show the cumulative response of y to a unit positive shock in x+ versus a unit negative shock in x− over the subsequent h periods. Diverging multiplier paths provide visual evidence of asymmetry; the width of the confidence bands indicates estimation uncertainty.

Sources

  1. Shin, Y., Yu, B., & Greenwood-Nimmo, M. (2014). Modelling asymmetric cointegration and dynamic multipliers in a nonlinear ARDL framework. In R. C. Sickles & W. C. Horrace (Eds.), Festschrift in Honor of Peter Schmidt (pp. 281–314). Springer. link ↗
  2. Becker, R., Enders, W., & Lee, J. (2006). A stationarity test in the presence of an unknown number of smooth breaks. Journal of Time Series Analysis, 27(3), 381–409. link ↗

How to cite this page

ScholarGate. (2026, June 3). Fourier Nonlinear Autoregressive Distributed Lag Model. ScholarGate. https://scholargate.app/en/econometrics/fourier-nardl

Related methods

Arellano-Bond GMM estimatorFourier ARDL Bounds TestFourier Engle-Granger cointegrationFourier Granger CausalityNonlinear ARDLVector Error Correction Model

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.

  • Arellano-Bond GMM estimatorEconometrics↔ compare
  • Fourier ARDL Bounds TestEconometrics↔ compare
  • Fourier Engle-Granger cointegrationEconometrics↔ compare
  • Fourier Granger CausalityEconometrics↔ compare
  • Nonlinear ARDLEconometrics↔ compare
  • Vector Error Correction ModelEconometrics↔ compare
Compare side by side →

Similar methods

Fourier ARDL Bounds TestNonlinear NARDLNonlinear ARDL bounds testNonlinear ARDLRobust NARDLStructural Break NARDLPanel NARDLNARDL Model

Related reference concepts

EconometricsSingle Equation Models • Single VariablesMultiple or Simultaneous Equation Models • Multiple VariablesMathematical and Quantitative MethodsFinancial EconometricsEconometric Modeling

Spotted an issue on this page? Report or suggest a fix →

ScholarGate — Fourier NARDL (Fourier Nonlinear Autoregressive Distributed Lag Model). Retrieved 2026-07-21 from https://scholargate.app/en/econometrics/fourier-nardl · Dataset: https://doi.org/10.5281/zenodo.20539026
Quick facts
Originator
Extension of Shin, Yu & Greenwood-Nimmo (2014) NARDL, incorporating Fourier terms from Becker, Enders & Lee (2006)
Year
2014–2020s
Type
Nonlinear cointegrating model with smooth break approximation
DataType
Time-series or panel time-series data
Subfamily
Econometrics / time series
Related methods
Arellano-Bond GMM estimatorFourier ARDL Bounds TestFourier Engle-Granger cointegrationFourier Granger CausalityNonlinear ARDLVector Error Correction Model
ScholarGate

A content-first reference library for research methods — what each one is, how it works, and where it comes from.

Open data (CC-BY)

Explore

  • Library
  • Search the library…
  • Browse by field
  • Fields
  • Journey
  • Compare
  • Which method?

Reference

  • Subjects
  • Atlas
  • Glossary
  • Methodology
  • Philosophy

Your tools

  • Bookshelf
  • Desk
  • Chat

Company

  • About
  • Pricing
  • Contact
  • Suggest a method

Entries are compiled from published sources for reference. Verifying the accuracy and suitability of any information for your own use remains your responsibility.

© 2026 ScholarGate · A research-method reference library
  • Privacy
  • Cookies
  • Terms
  • Delete account