A Unit Root Test Using a Fourier Series to Approximate Smooth Breaks
read more
Citations
Environmental Degradation in France: The Effects of FDI, Financial Development, and Energy Innovations
The flexible Fourier form and Dickey–Fuller type unit root tests
Financial development and environmental quality: The way forward
Public-private partnerships investment in energy as new determinant of CO2 emissions: The role of technological innovations in China
Decoupling CO2 emission and economic growth in China: Is there consistency in estimation results in analyzing environmental Kuznets curve?
References
Testing for a Unit Root in Time Series Regression
The Great Crash, the Oil Price Shock, and the Unit Root Hypothesis
Estimating and testing linear models with multiple structural changes
Further Evidence on Breaking Trend Functions in Macroeconomic Variables.
Unit Root Tests ARMA Models with Data Dependent Methods for the Selection of the Truncation Lag
Related Papers (5)
Frequently Asked Questions (13)
Q2. What is the critical value of LM(3) for the null hypothesis?
for T = 500, at the 5% significance level, the critical valueof τLM(3) for the null hypothesis that = 0 is 5.43 if frequencies 1, 2 and 3 are used in the estimating equation.
Q3. What is the problem with the power of the test?
The problem is that the power of the test diminishes rapidly as additional frequencycomponents are added to the estimating equation.
Q4. How do the authors obtain the time-varying mean of the yt sequence?
The time-varying mean of the {yt} sequence can be obtained by dividing the Fourierintercept [i.e., 0.160 + 0.080 sin(2kt/T) + 0.069 cos(2kt/T)] by one minus the sum of theautoregressive coefficients ( = 0.232).
Q5. How many Monte Carlo replications are rejected?
At the 5% significance level, the test rejects the null hypothesis of a unit root (i.e., = 1) in 4.9 % of the 20,000 Monte Carlo replications.
Q6. How do the authors determine k from the F-statistic?
To utilize their tests, wefirst obtain k̂ from (14) by minimizing the SSR and applying the F-test with F( k̂ ) to examine whether a nonlinear trend exists.
Q7. What is the simplest way to select a single k?
A data-driven method of selecting a single kA completely agnostic approach to the problem of detecting breaks is to select k usingpurely statistical means.
Q8. what is the fstatistic for a nonlinear trend?
In order to develop a test with a null hypothesis of linearity against the alternative of a nonlinear trend with given frequency k, the authors can also use the following Fstatistic:F(k)= 0 1 1 ( ( )) / 2 ( ) /( ) SSR SSR k SSR k T q (14)Here, SSR1(k) denotes the sum of squared residuals (SSR) from equation (10), q is thenumber of regressors, and SSR0 denotes the SSR from the regression without the trigonometric terms.
Q9. How many times did the null hypothesis of a unit root be rejected?
for n = 2 and = 0.9, the null hypothesis of a unit root was correctly rejected in 22.1%and 96% of the trials when T = 200 and 500, respectively.
Q10. What is the effect of ignoring a nonlinear trend?
Effects of Ignoring Nonlinear TrendsLemma 3 above indicates that ignoring a nonlinear trend affects the performance of alinear unit-root test under the null and alternative hypotheses.
Q11. How many squared residuals were obtained using only an intercept and trend?
The sum of squared residuals (SSR = 302.62) was about one third of thatobtained using only an intercept and trend (SSR = 896.40).
Q12. What are the likely associations of the higher frequency components of a series?
Becker et al. also show that the higher frequency components of a series are most likely to be associated with stochastic parameter variation.
Q13. What is the effect of ignoring a structural break in the DF test?
Perron (1989) earlier suggested that there will be a bias against rejecting a false unit root if an existing structural break is ignored in the usual DF test.