Fourier Hausman Test
Fourier Flexible Form Hausman Endogeneity Test · Also known as: Fourier-Hausman endogeneity test, Fourier augmented Hausman test, nonlinear Hausman test, flexible Hausman specification test
The Fourier Hausman test extends the classical Hausman endogeneity test by augmenting the regression with Fourier trigonometric terms — sines and cosines of time — so that the test remains valid even when the data-generating process contains smooth structural breaks or gradual nonlinearities that conventional linear specifications miss.
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 Hausman test when you want to test for endogeneity in a time-series or panel equation and you suspect that the relationship may have undergone gradual or smooth structural shifts — for example around financial crises, policy reforms, or slow technological transitions. It is preferable to the plain Hausman test whenever standard unit-root or specification pretests suggest the presence of smooth breaks. Do not apply it when the structural break is abrupt and well-dated (use a Chow or dummy-augmented Hausman instead), when the sample is very short (the Fourier terms consume degrees of freedom), or when no valid instruments are available (the test still requires instruments).
Strengths & limitations
- Robust to smooth structural breaks and gradual nonlinearities that invalidate standard linear Hausman tests.
- Does not require the researcher to pre-specify the break date or the functional form of the transition.
- Fourier terms are globally smooth and orthogonal, making them computationally tractable and interpretable.
- Retains the familiar chi-squared asymptotic distribution of the Hausman statistic under the null.
- Applicable in both single-equation time-series and panel data contexts.
- Requires at least one valid external instrument; the test provides no remedy if instruments are weak or absent.
- Adding Fourier terms reduces degrees of freedom, which can be costly in short samples (T < 60).
- Frequency selection via BIC is imperfect and adds a pre-test step that can inflate size in finite samples.
- Power may be lower than the plain Hausman test when the true model is genuinely linear with no structural breaks.
Frequently asked
How is the Fourier Hausman test different from the standard Hausman test?
The standard Hausman test compares OLS and IV estimates in a linear model; if the model omits a smooth structural break, the test may falsely flag endogeneity. The Fourier variant adds sine and cosine terms to control for gradual nonlinearities, so the endogeneity verdict is no longer confounded by misspecified trend behaviour.
How do I choose the number of Fourier frequencies?
Estimate the augmented model for k = 1, 2, … up to a small maximum (typically 3–5) and select the frequency that minimises BIC or AIC. Only one or two frequencies are usually sufficient; adding more quickly exhausts degrees of freedom in typical macro time series.
What instruments do I need?
The same instruments as in any IV or 2SLS application: variables that are correlated with the endogenous regressor but uncorrelated with the structural error, even after Fourier augmentation. The Fourier terms themselves are not instruments; they are controls in both stages.
What should I do if the test rejects exogeneity?
Proceed with IV or GMM estimation within the Fourier-augmented specification. The Fourier-augmented IV estimator is already available from the first stage; use it with HAC standard errors if serial correlation is suspected.
Can the Fourier Hausman test be applied to panel data?
Yes. In a panel setting the Fourier terms are typically added as common time effects or as individual-specific smooth trends, and the Hausman comparison is made within a fixed- or random-effects IV framework.
Sources
- Christopoulos, D. K., & Leon-Ledesma, M. A. (2004). Current account sustainability in the US: What do we really know about it? Journal of International Money and Finance, 23(5), 821–840. DOI: 10.2139/ssrn.596862 ↗
- Gallant, A. R. (1981). On the bias in flexible functional forms and an essentially unbiased form: The Fourier flexible form. Journal of Econometrics, 15(2), 211–245. DOI: 10.1016/0304-4076(81)90115-9 ↗
How to cite this page
ScholarGate. (2026, June 3). Fourier Flexible Form Hausman Endogeneity Test. ScholarGate. https://scholargate.app/en/econometrics/fourier-hausman-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.
- 2SLS RegressionEconometrics↔ compare
- Cointegration TestEconometrics↔ compare
- Granger CausalityEconometrics↔ compare
- Hausman TestEconometrics↔ compare
- Instrumental Variables in Health ResearchHealth Economics↔ compare